nature neuroscience volume 15 | number 4 | APrIl 2012 505

n e w s a n d v i e w s

8. Ungless, M.A. et al. Science 303, 2040–2042 (2004).

9. Schultz, W. et al. Science 275, 1593–1599 (1997).10. Bolam, J.P. & Smith, Y. Brain Res. 529, 57–78

(1990).11. Tepper, J.M. & Lee, C.R. Prog. Brain Res. 160,

189–208 (2007).12. Kitai, S.T. et al. Curr. Opin. Neurobiol. 9, 690–697

(1999).13. Johansen, J.P. & Fields, H.L. Nat. Neurosci. 7,

398–403 (2004).14. Jhou, T.C. et al. Neuron 61, 786–800 (2009).15. Borgius, L. et al. Mol. Cell. Neurosci. 45, 245–257

(2010).

COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests.

1. Schultz, W. J. Neurophysiol. 80, 1–27 (1998).2. Wise, R.A. Nat. Rev. Neurosci. 5, 483–494 (2004).3. Bromberg-Martin, E.S. et al. Neuron 68, 815–834

(2010).4. Brown, M.T.C. et al. J. Neurosci. 29, 2915–2925

(2009).5. Matsumoto, M. & Hikosaka, O. Nature 459, 837–841

(2009).6. Brischoux, F. et al. Proc. Natl. Acad. Sci. USA 106,

4894–4899 (2009).7. Henny, P. et al. Nat. Neurosci. 15, 613–619 (2012).

Henny et al.7 elegantly demonstrate how structural differences can give rise to func-tionally heterogeneous neuronal activity. By applying similar techniques to other regions, we can begin to address questions about which early neuroanatomists such as Ramón y Cajal could do no more than specu-late. Indeed, further research will continue to illuminate the complex interplay between form and function in the architecture of the brain.

family proteins, which bind and sequester the BH3-only proteins3,4. Other proteins in the Bcl-2 family (so named because they were first identified as being overexpressed in B-cell

Raymond A. Swanson is in the Neurology and

Rehabilitation Service, San Francisco Veterans

Affairs Medical Center, and the Department

of Neurology, University of California, San

Francisco, San Francisco, California, USA.

e-mail: raymond.swanson@ucsf.edu

one cell’s poison is another cell’s cureRaymond A Swanson

A drug that promotes cell death in cancer cells prevents cell death in post-ischemic neurons. These contrasting effects stem from the distinct actions of the anti-apoptotic protein Bcl-xL and its cleavage product, DN-Bcl-xL.

The cancer chemotherapeutic agent ABT-737 acts by promoting apoptosis, a form of pro-grammed cell death. In this issue, however, Ofengeim et al.1 find that ABT-737 can also prevent neuronal death after brain ischemia-reperfusion. How does an intervention that promotes cell death in one context prevent cell death in another? The authors resolve this paradox by finding that ischemic injury to neurons results in the cleavage of the anti-apoptotic protein Bcl-xL to a fragment (∆N-Bcl-xL) with pro-apoptotic properties, and that these pro-apoptotic effects are negated by ABT-737.

Under normal circumstances, apoptosis serves as a mechanism for the orderly removal of unneeded and dysfunctional cells2,3. Defects in apoptosis occur in many types of cancer, and targeted promotion of apoptosis is now a promising approach to cancer therapy4,5. Apoptosis can be triggered by several signals, but a common pathway that is activated by many of these signals is the oligomerization of the pro-apoptotic proteins Bak and Bax on the mitochondrial outer membrane (Fig. 1). Bak and Bax oligomerize in the presence of ‘BH3-only’ proteins to form large- conductance pores in the membrane. The resultant egress of cytochrome c and other mitochondrial proteins to the cytosol activates a cascade of proteolytic enzymes (termed caspases) that cause cell death by a series of steps

terminating in nuclear fragmentation and DNA degradation2,3.

The oligomerization of Bak and Bax is nor-mally held in check by Bcl-2 and other Bcl-2

Celldeath

∆N-Bcl-xL

Neuronalischemia–induced

apoptosis

Bax Bak Bcl-xL

Bcl-xLBH3

ABT-737

Inactivecaspase

Activecaspase

Bax Bak Bcl xxLax Bak Bcl-x

BH3

ABT 737

Basalstate

ABT-737–induced

apoptosis

∆N Bc

Celldeath

a b

Inactivecaspase

Activecaspase

Celldeath

Protectionby ABT-737

ABT-737

Inactivecaspase

Activecaspase

Figure 1 Contrasting effects of ABT-737 in cancer cells and post-ischemic neurons. Details of subcellular localization and protein-protein interactions are for illustrative purposes only. (a) Top: in the basal state, Bax and Bak proteins are monomeric and BH3-only proteins are bound by Bcl-xL. Cytochrome c is retained in the mitochondria. Bottom: in cancer cells, ABT-737 promotes pore formation and apoptotic cell death by binding to Bcl-xL, causing it to dissociate from BH3-only proteins. The unsequestered BH3-only proteins induce oligomerization of Bax and Bak, leading to large pore formation. Cytochrome c enters the cytosol through the pore and triggers caspase activation, leading to cell death. (b) Top: ischemia-reperfusion induces mitochondrial pore formation in neurons, which results in caspase-mediated Bcl-xL cleavage to ∆N-Bcl-xL. ∆N-Bcl-xL also forms pores, further stimulating caspases and promoting cell death. Bottom: ABT-737 administered after ischemia reperfusion blocks Bcl-xL cleavage to ∆N-Bcl-xL (dotted arrow) and independently, blocks ∆N-Bcl-xL–induced pore formation.

Kat

ie V

icar

i

npg

© 2

012

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.

506 volume 15 | number 4 | APrIl 2012 nature neuroscience

n e w s a n d v i e w s

be necessary for this to be clinically useful. Several therapeutic agents have been shown to be effective in reducing stroke size when given within a few hours of ischemia, but they lacked effect when tested in clinical trials using longer time-to-treat intervals14,15. In addi-tion, caspase-mediated cell death may have a larger role in the brief (10-min) brain ischemia used in the authors’ model than in the clini-cally more common setting of cerebral artery occlusion11. Future studies using other animal models and with longer post-ischemic time-to-treat intervals could more fully assess the promise of this approach for treating stroke.

COMPETING FINANCIAL INTERESTSThe author declares no competing financial interests.

1. Ofengeim, D. et al. Nat. Neurosci. 15, 574–580 (2012).

2. Kerr, J.F., Wyllie, A.H. & Currie, A.R. Br. J. Cancer 26, 239–257 (1972).

3. Martinou, J.C. & Youle, R.J. Dev. Cell 21, 92–101 (2011).

4. Kang, M.H. & Reynolds, C.P. Clin. Cancer Res. 15, 1126–1132 (2009).

5. Reed, J.C. & Pellecchia, M. Blood 106, 408–418 (2005).

6. Tsujimoto, Y., Finger, L.R., Yunis, J., Nowell, P.C. & Croce, C.M. Science 226, 1097–1099 (1984).

7. Oltersdorf, T. et al. Nature 435, 677–681 (2005).8. Zhang, F., Yin, W. & Chen, J. Neurol. Res. 26,

835–845 (2004). 9. MacManus, J.P., Buchan, A.M., Hill, I.E., Rasquinha, I. &

Preston, E. Neurosci. Lett. 164, 89–92 (1993). 10. Moskowitz, M.A., Lo, E.H. & Iadecola, C. Neuron 67,

181–198 (2010). 11. Lipton, P. Physiol. Rev. 79, 1431–1568 (1999).12. Fujita, N., Nagahashi, A., Nagashima, K., Rokudai, S. &

Tsuruo, T. Oncogene 17, 1295–1304 (1998). 13. Krajewska, M. et al. Cell Death Differ. 9, 145–157

(2002). 14. Ginsberg, M.D. Stroke 40, S111–S114 (2009). 15. Stroke Progress Review Group. Final report of the

Stroke Progress Review Group – January 2012. National Institute of Neurological Disorders and Stroke <http://www.ninds.nih.gov/find_people/ groups/stroke_prg/01-2012-stroke-prg-report.htm> (2012).

lymphomas6) include Bcl-xL and Bcl-w. The chemotherapeutic agent ABT-737 promotes apoptosis in cancer cells by displacing BH3-only proteins from Bcl-2 family proteins, thereby allowing the BH3-only proteins to facilitate Bak and Bax oligomerization4,7.

The same machinery that regulates apop-totic cell death in cancer cells is also engaged in neurons after ischemia, and contributes to ischemic neuronal death8–11. Thus, the obser-vation by Ofengeim et al.1 that the apoptosis-promoting agent ABT-737 reduced, rather than increased, ischemic brain injury is un expected. The key to this observation lies in the fact that the cleavage of the C-terminal portion of the Bcl-2 family protein Bcl-xL produces ∆N-Bcl-xL fragments with pro-apoptotic effects12. The authors found that ∆N-Bcl-xL was generated in neurons shortly after ischemia reperfusion and that ischemia-induced mitochondrial pore formation was blocked by ABT-737. Moreover, ∆N-Bcl-xL infused directly into neurons similarly induced the formation of an outer mitochondrial membrane pore and egress of cytochrome c, and these effects of ∆N-Bcl-xL were likewise blocked by co-infusion of ABT-737. In a transgenic mouse strain expressing a modified Bcl-xL protein lacking the caspase cleavage site, as in wild-type mice treated with ABT-737, there was reduced opening of mito-chondrial membrane pores and reduced neu-ronal death after ischemia reperfusion.

The effect of ABT-737 on neuronal mito-chondrial pore formation and ischemic injury is clearly demonstrated in these studies, but the mechanism by which ABT-737 exerts these effects remains less clear. The issue is compli-cated by the fact that Bcl-xL is both a regulator of caspase activation and a substrate for caspase cleavage (Fig. 1). As Ofengeim et al.1 note, this

suggests at least two possible mechanisms by which ABT-737 may affect neuronal survival: by binding to ∆N-Bcl-xL, thereby blocking its effect on mitochondrial pore formation, and by binding to full-length Bcl-xL (or other Bcl-2 family mem-bers), thereby preventing its cleavage and inac-tivation. The authors present evidence for both these processes, although how ∆N-Bcl-xL induces mitochondrial pore formation and how ABT-737 prevents this process remain uncertain.

The question also remains as to why the anti-apoptotic effects of ABT-737 trump the expected pro-apoptotic effects of this drug in post-ischemic neurons, but not in cancer cells. This question is not addressed by Ofengeim et al.1, but the relatively high Bcl-xL expres-sion in brain13, and differing levels of signal-ing for Bak and Bax oligomerization in cancer cells and post-ischemic neurons, may contri-bute. Alternatively, the reduced formation of large mitochondrial pores in the presence of ABT-737 may favorably affect mitochondrial bio-energetic function after ischemia reperfu-sion, independent of any other anti-apoptotic effects. Unlike mitochondria in other cell types, most mitochondria in neurons are located in cell processes that are distant from the nucleus and are therefore unable to participate in the cell execution phase of apoptosis. Thus, the pro-survival effect of ABT-737 in neurons may be a uniquely neuronal phenomenon.

Can this approach be used to treat brain ischemia? Perhaps, but a practical hurdle in the treatment of brain ischemia (stroke) is that the vast majority of affected individuals do not present for medical treatment until many hours after the onset of ischemia14. Ofengeim et al.1 found that ABT-737 can be effective when given as late as 1 h after ischemia reper fusion, but a much longer treatment window will

NATURE NEUROSCIENCE VOLUME 14 | NUMBER X | MONTH 2011 1

N E W S A N D V I E W S

Deep thinkingOne of the things that distinguish the picture on the right from a real toy ballerina is the feeling of depth that is evoked by a real object. Motion parallax is one such depth cue, when parts of the moving ballerina that are closer to the observer are more blurred than the parts that are further away. This is more obvious when a landscape is seen from the window of a moving train, when the foreground is more blurred than the background. Combined with other cues, such as binocular disparity (where the image received by each eye is slightly different and this difference depends on the distance of the object casting the image), information from motion parallax helps us to decide which things are near and which are far away.

On page 636 of this issue, Andrew Welchman and his colleagues at the University of Birmingham used functional imaging while testing their observers’ perception to determine how these depth cues are combined by the brain. Although it is known that people are very good at combining these cues, there are many ways that these cues might be used by the brain. Information from each cue might be extracted independently, in separate parts of the cortex, or this information might be fused into a single measure.

To test these different possibilities, the authors showed subjects moving or stationary dot patterns that evoked a feeling of depth because of binocular disparity, motion parallax or both. They then looked for brain areas that showed the characteristic signature of extracting information from combining both the cues, rather than extracting information from each independently. They found that the visual cortical area V3B/KO fit this characteristic profile.

Charvy Narain

npg

© 2

012

Nat

ure

Am

eric

a, In

c. A

ll rig

hts

rese

rved

.