npg
© 2014 Nature America, Inc. All rights reserved.
news and views likely to report perceiving or not perceiving the stimulus). There were striking differences in heartbeat-evoked responses that preceded those trials during which the target was perceived compared with those in which it was not seen. These differences predominantly localized to two cortical regions, right posterior parietal and ventromedial prefrontal cortex. The authors carefully eliminated a number of possible confounding factors, including general cortical excitability and specific measures of bodily arousal, as likely explanations of the differences. One of the important ways that this work extends our understanding is by demonstrating the importance of brain-body interactions in apparently ‘cold’ cognition. The visual stimulus used by Park et al.3 had no intrinsic emotional value and performance on the task was not rewarded directly. Yet the brain’s response to its body before the appearance of the stimulus was significant in guiding its eventual response to the stimulus itself. Given the arguments of James and his successors, it seems obvious that the state of the body, or the brain’s response to that state, in ‘hot’ (emotionally valenced) cognition might have substantial effects on perception and action, and such effects have been elegantly demonstrated11,12. But it is less immediately apparent why bodily states (or the brain’s interpretation of such states) should have such a marked effect on the apparently arbitrary visual perception task chosen by Park et al.3. The authors speculate that successful detection depends on an enhanced subjective feeling of the self at the time of the stimulus that might be provided by the enhanced neural responses to prestimulus heartbeats that they observed. This is reminiscent of theories of consciousness that emphasize embodiment as a precursor to the experience of subjective states14.
There are other possible interpretations of these striking results. For example, although the authors measured many bodily parameters, it is impossible in principle to rule out the possibility that there might be some difference in the physiology of those heartbeats (or associated with those heartbeats) that preceded correct detections rather than misses. This alternative is important, as it implies that the critical difference in cerebral responses to heartbeats that precede the stimulus might not be spontaneous, but is instead a reflection of an interaction between the stimulus and some aspect of the body’s state. For example, one theoretical possibility is that weaker than expected heartbeats might provoke greater heartbeat-evoked responses, but allow more sensitive performance at visual detection as a result of reduced physical motion of the body or the less distracting effects of such beats. Detecting such physical effects may be challenging and would require more invasive measures. Are the findings reported by Park et al.3 relevant to the observation that the heartbeat is tracked over wide areas of cortex 2? It is tempting to speculate that they might be; if information about the heartbeat is relevant to the psychophysically pure, but fundamentally humdrum, visual grating detection task performed by the participants in this study, it may turn out to be important in many other domains of perception and action. On the other hand, it is interesting that the critical differences in heartbeatevoked responses identified by Park et al.3 localized to brain regions not conventionally recognized as showing task-relevant activity during visual detection. Instead, these differences were seen in areas more frequently associated with ‘default mode’ activity 15
(Fig. 1), so perhaps these regions will ultimately be shown to integrate viscerally related responses into performance on a variety of tasks. Whatever the precise explanation and interpretation, these findings make a substantial contribution to the decades-old debate over the meaning of spontaneous activity: they suggest that at least some aspects of such activity are best regarded not as spontaneous but as differential responses to visceral states. It seems that, although the seat of sensation is not the heart (as Aristotle supposed), it might reasonably be argued that, in the light of this evidence, it is a crucial component of the source. Given how seriously the brain appears to be to taking the heartbeat, perhaps we as experimenters need to do so as well. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Gross, C.G. Neuroscientist 1, 245–250 (1995). 2. Kern, M., Aertsen, A., Schulze-Bonhage, A. & Ball, T. Neuroimage 81, 178–190 (2013). 3. Park, H.-D., Correia, S., Ducorps, A. & Tallon-Baudry, C. Nat. Neurosci. 17, 612–618 (2014). 4. Arieli, A., Sterkin, A., Grinvald, A. & Aertsen, A. Science 273, 1868–1871 (1996). 5. Boly, M. et al. Proc. Natl. Acad. Sci. USA 104, 12187–12192 (2007). 6. Fox, M.D., Snyder, A.Z., Vincent, J.L. & Raichle, M.E. Neuron 56, 171–184 (2007). 7. Angell, J.R. & Thompson, H.B. Psychol. Rev. 6, 32–69 (1899). 8. Cacioppo, J.T. J. Pers. Soc. Psychol. 37, 489 (1979). 9. Velden, M. & Juris, M. Psychophysiology 12, 685–692 (1975). 10. Walker, B.B. & Sandman, C.A. Psychophysiology 19, 520–527 (1982). 11. Gray, M.A., Rylander, K., Harrison, N.A., Wallin, B.G. & Critchley, H.D. J. Neurosci. 29, 1817–1825 (2009). 12. Gray, M.A. et al. Emotion 12, 180–191 (2012). 13. Schandry, R., Sparrer, B. & Weitkunat, R. Int. J. Neurosci. 30, 261–275 (1986). 14. Damasio, A.R. The Feeling of What Happens: Body and Emotion in the Making of Consciousness (Harcourt Brace, New York, 1999). 15. Gusnard, D.A., Raichle, M.E. & Raichle, M.E. Nat. Rev. Neurosci. 2, 685–694 (2001).
So many progenitors, so little myelin Steven A Goldman & Joana Osorio CNS white matter injury may cause sustained demyelination despite the persistence of oligodendrocyte progenitor cells (OPCs). A study suggests that dysregulated Wnt signaling disrupts self-renewal to yield OPC maturation arrest. Demyelinating injuries, as seen in multiple sclerosis, white matter stroke and many forms of cerebral palsy, are generally characterized Steven A. Goldman and Joana Osorio are in the Center for Translational Neuromedicine and the Department of Neurology of the University of Rochester Medical Center, Rochester, New York, USA. e-mail: [email protected]
by limited remyelination, and recurrent episodes of demyelination are especially bereft of recovery potential. Yet these same myelindeficient foci often retain ample numbers of precursor cells with the potential to become myelinating cells, namely oligodendrocyteastrocyte glial progenitor cells (OPCs), and seemingly arrested premyelinogenic oligodendroglia1–3. This paradoxical failure in
nature neuroscience volume 17 | number 4 | april 2014
remyelination despite the persistence of available progenitors has been the focus of recent research as part of a broader effort to improve remyelination competence. Although studies have made headway toward identifying triggers for oligodendrocyte differentiation and myelination from endogenous progenitors4,5, none have cracked the fundamental problem of why the injury environment is so often 483
news and views a
Wnt Frizzled PTPRZ1
Cytoplasm
Axin GSK3β APC
β-Cat
(–)
Regulated OPC self-renewal
Proteasome
β-Cat
Wnt-activated genes
TCF
Nucleus
Hypoxic ischemic injury
b
Wnt Frizzled PTPRZ1
npg
© 2014 Nature America, Inc. All rights reserved.
Cytoplasm
Axin GSK3β APC
β-Cat (–)
Dysregulated self-renewal: OPC differentiation failure Proteasome
TCF
(+)
β-Cat LEF1
Wnt-activated genes
(+) (+
L LEF1 SP5 S
SP1
R RNF43, DUSP4, ETS2, etc. Nucleus
Figure 1 OPC differentiation competence is diminished by developmental white matter injury. (a) The homeostatic self-renewal of OPCs in healthy forebrain white matter is modulated through the regulated nuclear translocation of β-catenin, its binding to TCF-family transcription factors and the resultant transcription of TCF-regulated target genes. Extracellular signals that modulate this process include the Wnt proteins in their binding to Frizzled-family receptors, the activation of which downregulates glycogen synthase kinase 3 beta (GSK-3β) activity, which would otherwise target cytoplasmic β-catenin for proteosomal degradation. Concurrent self-renewal signals also include the receptor tyrosine phosphatase β/ζ (PTPRZ1), which acts in concert with GSK3β to tonically suppress catenin availability9. (b) Following hypoxic-ischemic injury to white matter OPCs, β-catenin–mediated TCF-dependent transcription increases and expression of its negative regulator APC falls. These cells mimic the dysregulated Wnt/βcatenin signaling noted in APC-deficient colon cancers, with which they share upregulation of Wntmodulated target genes. This tonically upregulated gene set includes the LEF1 transcription factor, which drives its own expression in a feedforward manner to sustain high levels of Wnt-associated gene expression. It also includes the SP/KLF-family transcription factor SP5, which sustains self-renewal competence and whose knockdown yields accelerated remyelination. The feedforward nature of this dysregulated LEF overexpression may render OPCs refractory to terminal oligodendrocytic differentiation, instead favoring persistence in the progenitor state. However, unlike the neoplastic expansion seen in APC-mutant or APC-deficient colon carcinomas, the sustained Wnt/β-catenin signal activation and diminished APC expression seen in OPCs following hypoxic-ischemic injury is associated with maturational arrest, not with facilitated cell division or tumorigenesis, thereby highlighting the effects of phenotype and local tissue context on the cellular outcomes of dysregulated Wnt signaling.
nonpermissive for myelinogenesis from these cells. Do disease-associated changes in the local gliovascular niche present a nonpermissive environment for oligoneogenesis? Do the resident OPCs senesce under the pressures of sustained mobilization? Do they differentiate instead as reactive astrocytes rather than oligodendrocytes? Do they self-renew but undergo a discrete block in differentiation or myelinogenic competence? Although evidence has arisen in support of each of 484
these possibilities, the last possibility— that OPC self-renewal might be effectively commandeered to prevent terminal differentiation—has become of interest in recent studies of pathways whose manipulation has permitted the prolongation of OPC self-renewal. Perhaps foremost among these pathways has been that of Wnt-signaled β-catenin–dependent transcription, which normally contributes to OPC expansion (Fig. 1a),
but when dysregulated is associated with tumorigenesis from OPCs6–10. In this issue of Nature Neuroscience, Fancy et al.11 studied the role of Wnt-dependent β-catenin signaling in neonatal white matter injury using a transgenic mouse model of constitutively dysregulated β-catenin signaling. In this model, adenomatous polyposis coli (APC), an inhibitor of β-catenin nuclear availability and therefore a repressor of β-catenin-dependent, T cell factor 1 (TCF1)-mediated induction of Wnt target genes, is conditionally deleted in OPCs expressing the oligodendroglial lineage transcription factor Olig2. The OPCs thereby generated exhibited upregulation of a host of Wnt-regulated transcripts. Not surprisingly, many of these overlapped with Wntdependent transcripts that have already been recognized as being overexpressed in many colon cancers, which can be associated with disinhibited Wnt-driven gene expression due to APC mutation or deletion. But whereas the disinhibition of Wnt signaling associated with APC loss of function leads to carcinoma in colonic epithelial cells12,13, Fancy et al.11 noted neither neoplastic expansion nor malignant transformation in OPCs. On the contrary, they found that APC-null OPCs, despite their upregulated expression of Wntregulated transcripts, exhibited no change in expansion kinetics, but instead manifested maturation arrest and lost myelination competence. Moreover, their upregulation of the β-catenin binding partner LEF1 resulted in a feedforward self-induction of LEF1 transcription, which raised the effective gain, or ‘tone’, of Wnt signaling, in these OPCs, thereby stably potentiating β-catenin–dependent signaling in a self-perpetuating fashion (Fig. 1b). Fancy et al.11 then assessed the downstream targets of this process, recognizing that these molecules might provide valuable targets for modulating OPC differentiation and myelination, and therefore might offer feasible targets for rational drug discovery. Focusing on the transcript most highly differentially expressed by APC-null OPCs, the SP/KLF-family transcription factor SP5, the authors found that SP5 expression rose sharply with the overexpression of Wntregulated transcripts and that its overexpression was sufficient to suppress myelin basic protein expression by OPCs. Accordingly, SP5-null animals manifested potentiated oligodendrocytic differentiation and accelerated remyelination following demyelinating injury, suggesting the value of SP5 as a target for functional ablation in settings of dysregulated OPC self-renewal and arrested differentiation.
volume 17 | number 4 | april 2014 nature neuroscience
npg
© 2014 Nature America, Inc. All rights reserved.
news and views Yet many aspects of OPC biology differ fundamentally between mice and humans, so much so that one cannot assume that Wnt-regulated processes yield analogous cellular outcomes in mouse and human OPCs14. To address this concern, Fancy et al.11 assessed the differential expression of these Wnt-regulated transcripts in tissue samples from infants with severe hypoxic-ischemic encephalopathy. Perinatal hypoxic ischemia is a major cause of cerebral palsy, and although most presentations of cerebral palsy involve white matter loss, the contribution to its etiology of frank OPC loss, as opposed to dysfunction and maturational arrest, has been controversial2,15. Fancy et al.11 found that resident OPCs in the white matter lesions of these children shared remarkable molecular similarities with those derived from APC-null mice: they had no detectable APC and hence manifested derepressed Wnt signaling, with high levels of LEF1 and SP5, as well as other Wnt-driven, APC-deficient colon cancer–associated products that included v-ets avian erythroblastosis virus E26 oncogene homolog 2 (ETS2), dual specificity phosphatase 4 (DUSP4) and ring finger protein 43 (RFP43). Lest these expression patterns be considered transient, the authors describe bipolar RNF43+ OPCs in a 12-year-old child with cerebral palsy, whose perinatal hypoxic-ischemic event had presumably yielded not only a disabling degree of white matter loss, but also a functionally
insufficient OPC population whose maturational arrest had in no way diminished with the passage of time. It is the sustained nature of that maturational arrest that presents the greatest problems for those interested in mobilizing OPCs for therapeutic purposes. Can a transcription factor–based regulatory network sustain such an exquisitely balanced state of maturational arrest by a nominally self-renewing progenitor cell population without additional epigenetic changes to ensure the new status quo? If such changes do occur, will simple antagonism of Wnt-dependent pathways and downstream activators, such as SP5, be sufficient to reinitiate oligodendrocytic differentiation and myelination? Fancy et al.11 do not answer these questions, but they certainly do provide a window into how the virtual suspended animation of arrested, functionally compromised oligodendrocyte progenitors might be reversed. At the same time, the authors provide a model system in which those events involved in homeostatic maintenance may be decoupled from those involved in mitotic expansion, and the latter from oncogenesis and malignant transformation. As such, the worth of this study lies not only in its insights into the pathophysiology of hypoxic-ischemic white matter injury, but
also in its identification of processes by which we may now attempt to manipulate the mitotic expansion and myelination by endogenous progenitor cells across the entire spectrum of adult as well as pediatric myelin disorders, and how we might do so without the risk of concurrent oncogenesis. Rare indeed is it that such richness is offered in such a small package. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Franklin, R.J. & Ffrench-Constant, C. Nat. Rev. Neurosci. 9, 839–855 (2008). 2. Segovia, K.N. et al. Ann. Neurol. 63, 520–530 (2008). 3. Chang, A., Tourtellotte, W.W., Rudick, R. & Trapp, B.D. N. Engl. J. Med. 346, 165–173 (2002). 4. Huang, J.K. et al. Neurotherapeutics 8, 650–658 (2011). 5. Sim, F.J. et al. Ann. Neurol. 59, 763–779 (2006). 6. Chew, L.J. et al. J. Neurosci. 31, 13921–13935 (2011). 7. Fancy, S.P.J. et al. Nat. Neurosci. 14, 1009–1016 (2011). 8. Lang, J. et al. J. Neurosci. 33, 3113–3130 (2013). 9. McClain, C.R., Sim, F.J. & Goldman, S.A. J. Neurosci. 32, 15066–15075 (2012). 10. Fancy, S.P. et al. Genes Dev. 23, 1571–1585 (2009). 11. Fancy, S.P. et al. Nat. Neurosci. 17, 506–512 (2014). 12. Näthke, I. Nat. Rev. Cancer 6, 967–974 (2006). 13. Radtke, F. Science 307, 1904–1909 (2005). 14. Sim, F.J., Windrem, M.S. & Goldman, S.A. Neuron Glia Biol. 5, 45–55 (2009). 15. Buser, J.R. et al. Ann. Neurol. 71, 93–109 (2012).
Dark matter of the bulb Sasha Devore & Dmitry Rinberg A study now shows that granule cells deep in the olfactory bulb exhibit wildly different response dynamics depending on behavioral state, suggesting they could configure network changes across behavioral states. The olfactory system is under constant assault from an ever-changing mix of volatile molecules that waft through the air, impinging on the nasal mucosae with each new inhalation, yet the brain effortlessly extracts the relevant signals to guide behavior. The nervous system solves this task in part by dynamically tuning the olfactory bulb, the earliest olfactory processing network in the CNS, rendering it dependent on factors such as the wakefulness of the animal, the hedonic value of the incoming odorants and the recent history of odor stimulation (for review, see ref. 1). Flexible sensory information processing is thought to be mediated by extensive feedback projections Sasha Devore and Dmitry Rinberg are at the NYU Neuroscience Institute, New York University Langone Medical Center, New York, New York, USA. e-mail: [email protected]
to the olfactory bulb from cortex and subcortical neuromodulatory nuclei2–4. In particular, the granule cells, an extensive network of inhibitory neurons deep within the bulb, integrate inputs from nearly all of these centrifugal projections and are thought to orchestrate bulbar dynamics across behavioral states. Although they have been the subject of many theoretical models of bulbar processing5,6, relatively little is known empirically about the function of granule cells in vivo. A study by Cazakoff et al.7 in this issue of Nature Neuroscience opens a window into the deepest layers of the bulb by characterizing, for the first time, the odor- and breathing-related activity of individual granule cells in awake mice. Odor molecules in the air enter the nose by inhalation, and they are first detected by olfactory sensory neurons (OSNs) embedded in the nasal epithelium, each typically
nature neuroscience volume 17 | number 4 | april 2014
expressing a single olfactory receptor gene (selected from ~1,200 in rodent and ~300 in humans). In an impressive display of convergence, the axons from OSNs expressing the same receptor bundle together, forming distinctive glomerular structures in the outermost layers of the olfactory bulb (Fig. 1). Mitral and tufted (M/T) cells, the principal output neurons of the bulb, each send a primary dendrite into a single glomerulus. According to this basic wiring diagram, M/T cells are poised to convey information detected by a single receptor type to downstream cortical targets. However, extensive inter- and intraglomerular processing can occur via neurochemically and anatomically diverse networks of inhibitory neurons, which outnumber M/T cells by almost two orders of magnitude8. The responses of M/T cells to odors are far from simple: in awake animals, 485
© 2014 Nature America, Inc. All rights reserved.
news and views likely to report perceiving or not perceiving the stimulus). There were striking differences in heartbeat-evoked responses that preceded those trials during which the target was perceived compared with those in which it was not seen. These differences predominantly localized to two cortical regions, right posterior parietal and ventromedial prefrontal cortex. The authors carefully eliminated a number of possible confounding factors, including general cortical excitability and specific measures of bodily arousal, as likely explanations of the differences. One of the important ways that this work extends our understanding is by demonstrating the importance of brain-body interactions in apparently ‘cold’ cognition. The visual stimulus used by Park et al.3 had no intrinsic emotional value and performance on the task was not rewarded directly. Yet the brain’s response to its body before the appearance of the stimulus was significant in guiding its eventual response to the stimulus itself. Given the arguments of James and his successors, it seems obvious that the state of the body, or the brain’s response to that state, in ‘hot’ (emotionally valenced) cognition might have substantial effects on perception and action, and such effects have been elegantly demonstrated11,12. But it is less immediately apparent why bodily states (or the brain’s interpretation of such states) should have such a marked effect on the apparently arbitrary visual perception task chosen by Park et al.3. The authors speculate that successful detection depends on an enhanced subjective feeling of the self at the time of the stimulus that might be provided by the enhanced neural responses to prestimulus heartbeats that they observed. This is reminiscent of theories of consciousness that emphasize embodiment as a precursor to the experience of subjective states14.
There are other possible interpretations of these striking results. For example, although the authors measured many bodily parameters, it is impossible in principle to rule out the possibility that there might be some difference in the physiology of those heartbeats (or associated with those heartbeats) that preceded correct detections rather than misses. This alternative is important, as it implies that the critical difference in cerebral responses to heartbeats that precede the stimulus might not be spontaneous, but is instead a reflection of an interaction between the stimulus and some aspect of the body’s state. For example, one theoretical possibility is that weaker than expected heartbeats might provoke greater heartbeat-evoked responses, but allow more sensitive performance at visual detection as a result of reduced physical motion of the body or the less distracting effects of such beats. Detecting such physical effects may be challenging and would require more invasive measures. Are the findings reported by Park et al.3 relevant to the observation that the heartbeat is tracked over wide areas of cortex 2? It is tempting to speculate that they might be; if information about the heartbeat is relevant to the psychophysically pure, but fundamentally humdrum, visual grating detection task performed by the participants in this study, it may turn out to be important in many other domains of perception and action. On the other hand, it is interesting that the critical differences in heartbeatevoked responses identified by Park et al.3 localized to brain regions not conventionally recognized as showing task-relevant activity during visual detection. Instead, these differences were seen in areas more frequently associated with ‘default mode’ activity 15
(Fig. 1), so perhaps these regions will ultimately be shown to integrate viscerally related responses into performance on a variety of tasks. Whatever the precise explanation and interpretation, these findings make a substantial contribution to the decades-old debate over the meaning of spontaneous activity: they suggest that at least some aspects of such activity are best regarded not as spontaneous but as differential responses to visceral states. It seems that, although the seat of sensation is not the heart (as Aristotle supposed), it might reasonably be argued that, in the light of this evidence, it is a crucial component of the source. Given how seriously the brain appears to be to taking the heartbeat, perhaps we as experimenters need to do so as well. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Gross, C.G. Neuroscientist 1, 245–250 (1995). 2. Kern, M., Aertsen, A., Schulze-Bonhage, A. & Ball, T. Neuroimage 81, 178–190 (2013). 3. Park, H.-D., Correia, S., Ducorps, A. & Tallon-Baudry, C. Nat. Neurosci. 17, 612–618 (2014). 4. Arieli, A., Sterkin, A., Grinvald, A. & Aertsen, A. Science 273, 1868–1871 (1996). 5. Boly, M. et al. Proc. Natl. Acad. Sci. USA 104, 12187–12192 (2007). 6. Fox, M.D., Snyder, A.Z., Vincent, J.L. & Raichle, M.E. Neuron 56, 171–184 (2007). 7. Angell, J.R. & Thompson, H.B. Psychol. Rev. 6, 32–69 (1899). 8. Cacioppo, J.T. J. Pers. Soc. Psychol. 37, 489 (1979). 9. Velden, M. & Juris, M. Psychophysiology 12, 685–692 (1975). 10. Walker, B.B. & Sandman, C.A. Psychophysiology 19, 520–527 (1982). 11. Gray, M.A., Rylander, K., Harrison, N.A., Wallin, B.G. & Critchley, H.D. J. Neurosci. 29, 1817–1825 (2009). 12. Gray, M.A. et al. Emotion 12, 180–191 (2012). 13. Schandry, R., Sparrer, B. & Weitkunat, R. Int. J. Neurosci. 30, 261–275 (1986). 14. Damasio, A.R. The Feeling of What Happens: Body and Emotion in the Making of Consciousness (Harcourt Brace, New York, 1999). 15. Gusnard, D.A., Raichle, M.E. & Raichle, M.E. Nat. Rev. Neurosci. 2, 685–694 (2001).
So many progenitors, so little myelin Steven A Goldman & Joana Osorio CNS white matter injury may cause sustained demyelination despite the persistence of oligodendrocyte progenitor cells (OPCs). A study suggests that dysregulated Wnt signaling disrupts self-renewal to yield OPC maturation arrest. Demyelinating injuries, as seen in multiple sclerosis, white matter stroke and many forms of cerebral palsy, are generally characterized Steven A. Goldman and Joana Osorio are in the Center for Translational Neuromedicine and the Department of Neurology of the University of Rochester Medical Center, Rochester, New York, USA. e-mail: [email protected]
by limited remyelination, and recurrent episodes of demyelination are especially bereft of recovery potential. Yet these same myelindeficient foci often retain ample numbers of precursor cells with the potential to become myelinating cells, namely oligodendrocyteastrocyte glial progenitor cells (OPCs), and seemingly arrested premyelinogenic oligodendroglia1–3. This paradoxical failure in
nature neuroscience volume 17 | number 4 | april 2014
remyelination despite the persistence of available progenitors has been the focus of recent research as part of a broader effort to improve remyelination competence. Although studies have made headway toward identifying triggers for oligodendrocyte differentiation and myelination from endogenous progenitors4,5, none have cracked the fundamental problem of why the injury environment is so often 483
news and views a
Wnt Frizzled PTPRZ1
Cytoplasm
Axin GSK3β APC
β-Cat
(–)
Regulated OPC self-renewal
Proteasome
β-Cat
Wnt-activated genes
TCF
Nucleus
Hypoxic ischemic injury
b
Wnt Frizzled PTPRZ1
npg
© 2014 Nature America, Inc. All rights reserved.
Cytoplasm
Axin GSK3β APC
β-Cat (–)
Dysregulated self-renewal: OPC differentiation failure Proteasome
TCF
(+)
β-Cat LEF1
Wnt-activated genes
(+) (+
L LEF1 SP5 S
SP1
R RNF43, DUSP4, ETS2, etc. Nucleus
Figure 1 OPC differentiation competence is diminished by developmental white matter injury. (a) The homeostatic self-renewal of OPCs in healthy forebrain white matter is modulated through the regulated nuclear translocation of β-catenin, its binding to TCF-family transcription factors and the resultant transcription of TCF-regulated target genes. Extracellular signals that modulate this process include the Wnt proteins in their binding to Frizzled-family receptors, the activation of which downregulates glycogen synthase kinase 3 beta (GSK-3β) activity, which would otherwise target cytoplasmic β-catenin for proteosomal degradation. Concurrent self-renewal signals also include the receptor tyrosine phosphatase β/ζ (PTPRZ1), which acts in concert with GSK3β to tonically suppress catenin availability9. (b) Following hypoxic-ischemic injury to white matter OPCs, β-catenin–mediated TCF-dependent transcription increases and expression of its negative regulator APC falls. These cells mimic the dysregulated Wnt/βcatenin signaling noted in APC-deficient colon cancers, with which they share upregulation of Wntmodulated target genes. This tonically upregulated gene set includes the LEF1 transcription factor, which drives its own expression in a feedforward manner to sustain high levels of Wnt-associated gene expression. It also includes the SP/KLF-family transcription factor SP5, which sustains self-renewal competence and whose knockdown yields accelerated remyelination. The feedforward nature of this dysregulated LEF overexpression may render OPCs refractory to terminal oligodendrocytic differentiation, instead favoring persistence in the progenitor state. However, unlike the neoplastic expansion seen in APC-mutant or APC-deficient colon carcinomas, the sustained Wnt/β-catenin signal activation and diminished APC expression seen in OPCs following hypoxic-ischemic injury is associated with maturational arrest, not with facilitated cell division or tumorigenesis, thereby highlighting the effects of phenotype and local tissue context on the cellular outcomes of dysregulated Wnt signaling.
nonpermissive for myelinogenesis from these cells. Do disease-associated changes in the local gliovascular niche present a nonpermissive environment for oligoneogenesis? Do the resident OPCs senesce under the pressures of sustained mobilization? Do they differentiate instead as reactive astrocytes rather than oligodendrocytes? Do they self-renew but undergo a discrete block in differentiation or myelinogenic competence? Although evidence has arisen in support of each of 484
these possibilities, the last possibility— that OPC self-renewal might be effectively commandeered to prevent terminal differentiation—has become of interest in recent studies of pathways whose manipulation has permitted the prolongation of OPC self-renewal. Perhaps foremost among these pathways has been that of Wnt-signaled β-catenin–dependent transcription, which normally contributes to OPC expansion (Fig. 1a),
but when dysregulated is associated with tumorigenesis from OPCs6–10. In this issue of Nature Neuroscience, Fancy et al.11 studied the role of Wnt-dependent β-catenin signaling in neonatal white matter injury using a transgenic mouse model of constitutively dysregulated β-catenin signaling. In this model, adenomatous polyposis coli (APC), an inhibitor of β-catenin nuclear availability and therefore a repressor of β-catenin-dependent, T cell factor 1 (TCF1)-mediated induction of Wnt target genes, is conditionally deleted in OPCs expressing the oligodendroglial lineage transcription factor Olig2. The OPCs thereby generated exhibited upregulation of a host of Wnt-regulated transcripts. Not surprisingly, many of these overlapped with Wntdependent transcripts that have already been recognized as being overexpressed in many colon cancers, which can be associated with disinhibited Wnt-driven gene expression due to APC mutation or deletion. But whereas the disinhibition of Wnt signaling associated with APC loss of function leads to carcinoma in colonic epithelial cells12,13, Fancy et al.11 noted neither neoplastic expansion nor malignant transformation in OPCs. On the contrary, they found that APC-null OPCs, despite their upregulated expression of Wntregulated transcripts, exhibited no change in expansion kinetics, but instead manifested maturation arrest and lost myelination competence. Moreover, their upregulation of the β-catenin binding partner LEF1 resulted in a feedforward self-induction of LEF1 transcription, which raised the effective gain, or ‘tone’, of Wnt signaling, in these OPCs, thereby stably potentiating β-catenin–dependent signaling in a self-perpetuating fashion (Fig. 1b). Fancy et al.11 then assessed the downstream targets of this process, recognizing that these molecules might provide valuable targets for modulating OPC differentiation and myelination, and therefore might offer feasible targets for rational drug discovery. Focusing on the transcript most highly differentially expressed by APC-null OPCs, the SP/KLF-family transcription factor SP5, the authors found that SP5 expression rose sharply with the overexpression of Wntregulated transcripts and that its overexpression was sufficient to suppress myelin basic protein expression by OPCs. Accordingly, SP5-null animals manifested potentiated oligodendrocytic differentiation and accelerated remyelination following demyelinating injury, suggesting the value of SP5 as a target for functional ablation in settings of dysregulated OPC self-renewal and arrested differentiation.
volume 17 | number 4 | april 2014 nature neuroscience
npg
© 2014 Nature America, Inc. All rights reserved.
news and views Yet many aspects of OPC biology differ fundamentally between mice and humans, so much so that one cannot assume that Wnt-regulated processes yield analogous cellular outcomes in mouse and human OPCs14. To address this concern, Fancy et al.11 assessed the differential expression of these Wnt-regulated transcripts in tissue samples from infants with severe hypoxic-ischemic encephalopathy. Perinatal hypoxic ischemia is a major cause of cerebral palsy, and although most presentations of cerebral palsy involve white matter loss, the contribution to its etiology of frank OPC loss, as opposed to dysfunction and maturational arrest, has been controversial2,15. Fancy et al.11 found that resident OPCs in the white matter lesions of these children shared remarkable molecular similarities with those derived from APC-null mice: they had no detectable APC and hence manifested derepressed Wnt signaling, with high levels of LEF1 and SP5, as well as other Wnt-driven, APC-deficient colon cancer–associated products that included v-ets avian erythroblastosis virus E26 oncogene homolog 2 (ETS2), dual specificity phosphatase 4 (DUSP4) and ring finger protein 43 (RFP43). Lest these expression patterns be considered transient, the authors describe bipolar RNF43+ OPCs in a 12-year-old child with cerebral palsy, whose perinatal hypoxic-ischemic event had presumably yielded not only a disabling degree of white matter loss, but also a functionally
insufficient OPC population whose maturational arrest had in no way diminished with the passage of time. It is the sustained nature of that maturational arrest that presents the greatest problems for those interested in mobilizing OPCs for therapeutic purposes. Can a transcription factor–based regulatory network sustain such an exquisitely balanced state of maturational arrest by a nominally self-renewing progenitor cell population without additional epigenetic changes to ensure the new status quo? If such changes do occur, will simple antagonism of Wnt-dependent pathways and downstream activators, such as SP5, be sufficient to reinitiate oligodendrocytic differentiation and myelination? Fancy et al.11 do not answer these questions, but they certainly do provide a window into how the virtual suspended animation of arrested, functionally compromised oligodendrocyte progenitors might be reversed. At the same time, the authors provide a model system in which those events involved in homeostatic maintenance may be decoupled from those involved in mitotic expansion, and the latter from oncogenesis and malignant transformation. As such, the worth of this study lies not only in its insights into the pathophysiology of hypoxic-ischemic white matter injury, but
also in its identification of processes by which we may now attempt to manipulate the mitotic expansion and myelination by endogenous progenitor cells across the entire spectrum of adult as well as pediatric myelin disorders, and how we might do so without the risk of concurrent oncogenesis. Rare indeed is it that such richness is offered in such a small package. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. 1. Franklin, R.J. & Ffrench-Constant, C. Nat. Rev. Neurosci. 9, 839–855 (2008). 2. Segovia, K.N. et al. Ann. Neurol. 63, 520–530 (2008). 3. Chang, A., Tourtellotte, W.W., Rudick, R. & Trapp, B.D. N. Engl. J. Med. 346, 165–173 (2002). 4. Huang, J.K. et al. Neurotherapeutics 8, 650–658 (2011). 5. Sim, F.J. et al. Ann. Neurol. 59, 763–779 (2006). 6. Chew, L.J. et al. J. Neurosci. 31, 13921–13935 (2011). 7. Fancy, S.P.J. et al. Nat. Neurosci. 14, 1009–1016 (2011). 8. Lang, J. et al. J. Neurosci. 33, 3113–3130 (2013). 9. McClain, C.R., Sim, F.J. & Goldman, S.A. J. Neurosci. 32, 15066–15075 (2012). 10. Fancy, S.P. et al. Genes Dev. 23, 1571–1585 (2009). 11. Fancy, S.P. et al. Nat. Neurosci. 17, 506–512 (2014). 12. Näthke, I. Nat. Rev. Cancer 6, 967–974 (2006). 13. Radtke, F. Science 307, 1904–1909 (2005). 14. Sim, F.J., Windrem, M.S. & Goldman, S.A. Neuron Glia Biol. 5, 45–55 (2009). 15. Buser, J.R. et al. Ann. Neurol. 71, 93–109 (2012).
Dark matter of the bulb Sasha Devore & Dmitry Rinberg A study now shows that granule cells deep in the olfactory bulb exhibit wildly different response dynamics depending on behavioral state, suggesting they could configure network changes across behavioral states. The olfactory system is under constant assault from an ever-changing mix of volatile molecules that waft through the air, impinging on the nasal mucosae with each new inhalation, yet the brain effortlessly extracts the relevant signals to guide behavior. The nervous system solves this task in part by dynamically tuning the olfactory bulb, the earliest olfactory processing network in the CNS, rendering it dependent on factors such as the wakefulness of the animal, the hedonic value of the incoming odorants and the recent history of odor stimulation (for review, see ref. 1). Flexible sensory information processing is thought to be mediated by extensive feedback projections Sasha Devore and Dmitry Rinberg are at the NYU Neuroscience Institute, New York University Langone Medical Center, New York, New York, USA. e-mail: [email protected]
to the olfactory bulb from cortex and subcortical neuromodulatory nuclei2–4. In particular, the granule cells, an extensive network of inhibitory neurons deep within the bulb, integrate inputs from nearly all of these centrifugal projections and are thought to orchestrate bulbar dynamics across behavioral states. Although they have been the subject of many theoretical models of bulbar processing5,6, relatively little is known empirically about the function of granule cells in vivo. A study by Cazakoff et al.7 in this issue of Nature Neuroscience opens a window into the deepest layers of the bulb by characterizing, for the first time, the odor- and breathing-related activity of individual granule cells in awake mice. Odor molecules in the air enter the nose by inhalation, and they are first detected by olfactory sensory neurons (OSNs) embedded in the nasal epithelium, each typically
nature neuroscience volume 17 | number 4 | april 2014
expressing a single olfactory receptor gene (selected from ~1,200 in rodent and ~300 in humans). In an impressive display of convergence, the axons from OSNs expressing the same receptor bundle together, forming distinctive glomerular structures in the outermost layers of the olfactory bulb (Fig. 1). Mitral and tufted (M/T) cells, the principal output neurons of the bulb, each send a primary dendrite into a single glomerulus. According to this basic wiring diagram, M/T cells are poised to convey information detected by a single receptor type to downstream cortical targets. However, extensive inter- and intraglomerular processing can occur via neurochemically and anatomically diverse networks of inhibitory neurons, which outnumber M/T cells by almost two orders of magnitude8. The responses of M/T cells to odors are far from simple: in awake animals, 485
