news and views

nature neuroscience • volume 3 no 9 • september 2000 851

to develop technology that amplifiessignals at the launch end14. Thus, itseems that telecommunications engi-neers have begun, unwittingly, tomimic a trick used by neurons to com-pensate for attenuation. Neurons,however, are faced with an additionalproblem (not an issue for modernoptical cables), which is that signalsare filtered as they propagate. The highdensity of hyperpolarization-activatedchannels in CA1 dendrites provides anelegant solution to this problem bycompensating for this distance-depen-dent filtering. These findings providegreat insight into the process of synap-tic integration and dramatically nar-row the focus of how we think aboutneural network function. As we seek to

6. Jaffe, D. B. & Carnevale, N. T. J. Neurophysiol.82, 3268–3285 (1999).

7. Malenka, R. C. & Nicoll, R. A. Neuron 19,473–476 (1997).

8. Hoffman, D. A., Magee, J. C., Colbert, C.M. & Johnston, D. Nature 387, 869–875(1997).

9. Cash, S. & Yuste, R. Neuron 22, 383–394(1999).

10. Stuart, G., Spruston, N., Sakmann, B. & Häusser, M. Trends Neurosci. 20, 125–131(1997).

11. Linden, D. J. Neuron 22, 661–666 (1999).

12. Spruston, N., Stuart, G. & Häusser, M. inDendrites (eds. Stuart, G., Spruston, N. &Häusser, M.) 231–270 (Oxford Univ. Press,Oxford, 1999).

13. Magee, J. C. Nat. Neurosci. 2, 508–514(1999).

14. Hansen, P. Laser Focus World 34, 79–88(1998).

unravel the mechanisms of learningand memory in the hippocampus, wenow know that excitatory synapsesshould be treated equally regardingtheir strength and summation in thesoma, regardless of their position onthe dendritic tree.

1. Magee, J. C. & Cook, E. P. Nat. Neurosci. 3,895–903 (2000).

2. Bergano, N. S. Optics Photonics News 11,20–30 (2000).

3. Rall, W. in Handbook of Physiology: TheNervous System. Cellular Biology on NeuronsVol. 1 39–97 (Am. Phys. Soc., Bethesda,Maryland, 1977).

4. Magee, J. C. in Dendrites (eds. Stuart, G.,Spruston, N., Häusser, M.) 139–160(Oxford Univ. Press, Oxford, 1999).

5. Pettit, D. L. & Augustine, G. J. J. Neurophysiol. 84, 28–38 (2000).

The biological role of ciliary neu-rotrophic factor (CNTF) remains large-ly an enigma. It was first identified as atrophic factor in chick eye and nerveextracts that supported the survival ofciliary ganglionic neurons in vitro;mammalian CNTF was subsequentlypurified and cloned1,2. Intriguingly,CNTF was also found to support thesurvival of neurons in the CNS, includ-ing clinically important dopaminergicneurons in the substantia nigra as wellas motor neurons; these observationsgenerated hope that CNTF could beused as a therapeutic agent in neurode-generative diseases. The existence—andimportance—of another CNTF-like fac-tor, however, was suggested becauseCNTF itself is not a secreted proteinand, furthermore, because 2.3% of theJapanese population is homozygous fora null mutation of CNTF but neverthe-less shows no neurological deficits even

at advanced ages3. These and otherobservations suggested a rather restrict-ed role for CNTF, perhaps as a trophicfactor released after traumatic injury toperipheral nerves, and suggested that a‘CNTF II’ must exist that normally sup-ports the important biological functionsof the CNTF signaling pathway1,2.

Evidence for this elusive CNTF IIwas given a strong boost by a studycomparing the phenotypic conse-quences of disrupting CNTF versusCNTR receptor α subunit (CNTFRα)genes4. Although CNTF–/– mice appearlargely normal, CNTFRα gene disrup-tion is devastating, with deletions inmotor nuclei causing perinatal deathfrom the inability to suckle. In this issue,Elson and colleagues report the charac-terization of what promises to be thislong-sought second ligand for the CNTFreceptor5.

The relationship between ‘neuralcytokines’ like CNTF and hematopoiet-ic cytokines (typified by interleukin-6and granulocyte colony stimulating fac-tor6,7) helps to set the stage for under-

CNTF II, I presume?Steven S. Lesser and Donald C. Lo

Elson and colleagues report the identification of a new,secreted ligand for the ciliary neurotrophic factor receptor,which is likely to be important during development.

Fig. 1. Two principal transmembrane receptor subunits, gp130 and LIFRβ, form the core of abroad range of receptor complexes for neural and hematopoietic cytokines. The addition of athird ‘α’ subunit confers ligand selectivity to receptors for interleukin-6 (IL-6), CNTF and perhapsCT-1 (upper row), whereas an additional component that is part receptor–part ligand is requiredfor factor secretion and receptor activation in the cases of IL-12 and CLC–CLF (lower row).

The authors are in the Department ofNeurobiology, Box 3209, Duke UniversityMedical Center, Durham, North Carolina,27710, USA.email: lo@neuro.duke.edu orssless@neuro.duke.edu

© 2000 Nature America Inc. • http://neurosci.nature.com©

200

0 N

atu

re A

mer

ica

Inc.

• h

ttp

://n

euro

sci.n

atu

re.c

om

852 nature neuroscience • volume 3 no 9 • september 2000

news and views

standing the new ligand and its recep-tor. A key event for understanding theparallels between these signaling mole-cules was the discovery that cholinergicdifferentiation factor (CDF) is identicalto leukemia inhibitory factor (LIF).CDF was originally identified as asecreted activity that could induce phe-notypic switching in sympathetic neu-rons, whereas LIF was identified as afactor inducing the terminal differenti-ation of myeloid leukemia cells6. Suchfunctional parallels and overlapsbetween ‘neural cytokines’ andhematopoietic factors arise from theirstructural similarities at both the factorand receptor levels. CNTF turned out tobe a member of a broad family ofhematopoietic factors related in tertiarystructure7, and its primary receptor,CNTFRα , is structurally related to theα subunit of the receptor for inter-leukin-6 (ref. 8). Signal transductionstudies showed that the receptor com-plex for CNTF not only closely parallelsthat for interleukin-6, but also containstwo membrane-spanning subunits(gp130 and LIFRβ) that are shared by ahost of receptor complexes forhematopoietic cytokines1 (Fig. 1).Whereas CNTFRα confers selectivity forCNTF upon the receptor complex,gp130 and LIFRβ transduce ligandbinding into cellular responses via acti-vation of the JAK-STAT signaling path-way1,2. CNTFRα itself is not atransmembrane protein, but is anchoredextracellularly to the plasma membranevia a glycosylphosphatidylinositol (GPI)linkage. Without CNTFRα, the remain-ing two subunits constitute a receptorfor LIF, whereas two gp130 subunitsplus the interleukin-6 receptor α formthe functional receptor complex forinterleukin-6. This scheme of mixingand matching various α and β subunitsproduces a range of receptor complex-es from a relatively limited set of recep-tor components (Fig. 1). Because thetransmembrane subunits are commonto many receptor complexes, the down-stream effects of receptor activation bydifferent ligands are often similar if notidentical1,2.

The work reported by Elson and col-leagues5 in this issue stems from theidentification last year of a new mem-ber of the IL-6 class of cytokines, whichwas termed cardiotrophin-like cytokine(CLC) based on its homology to car-diotrophin-1, itself a relative of CNTF9,10;CLC was independently identified asnovel neurotrophin-1 (NNT-1)11. Previ-

lated as NR6) results in a phenotypewhere mice are unable to suckle15, rem-iniscent of CNTFRα knockouts. It willbe interesting to determine if CLF andCNTFRα gene disruption produce sim-ilar underlying deficits in the motorneuron populations in the brainstemand spinal cord. Finally, it will beimportant to positively identifyCLC–CLF as ‘CNTF II’—and to excludethe possibility of yet more ligands forthe CNTF receptor—by demonstratingthat a double knockout for CNTF andCLC/CLF encompasses the phenotyperesulting from CNTFRα gene disrup-tion.

Perhaps the discovery of this new fac-tor complex will rekindle efforts to findtherapeutic applications for CNTFreceptor signaling in nervous system diseases and disorders. Thus far,attempts to use CNTF as a therapeuticagent in the treatment of disorders suchas amyotropic lateral sclerosis have notmet with success, although CNTF is nowbeing used in clinical trials again for oneof its side effects, weight loss (RegeneronPharmaceuticals, http://graphics.regen-eron.com/research/researchAxokine.htm).Interestingly enough, this biologicalaction of CNTF probably arises fromparallels in its signal transduction cas-cade with yet another cytokine signalingsystem, that for leptin, but that is thesubject for a completely different Newsand Views…

1. Ip, N. Y. & Yancopoulos, G. D. Annu. Rev.Neurosci. 19, 491–515 (1996).

2. Segal, R. A. & Greenberg, M. E. Annu. Rev.Neurosci. 19, 463–489 (1996).

3. Takahashi, R. et al. Nat. Genet. 7, 79–84(1994).

4. DeChiara, T. M. et al. Cell 83, 313–322(1995).

5. Elson, G. C. A. et al. Nat. Neurosci. 3,867–872 (2000).

6. Patterson, P. H. & Nawa, H. Neuron 10(Suppl.), 123–137 (1993).

7. Bazan, J. F. Neuron 7, 197–208 (1991).

8. Davis, S. et al. Science 253, 59–63 (1991).

9. Shi, Y. et al. Biochem. Biophys. Res. Comm.262, 132–138 (1999).

10. Pennica, D. et al. Proc. Natl. Acad. Sci. USA92, 1142–1146 (1995).

11. Senaldi, G. et al. Proc. Natl. Acad. Sci. USA96, 11458–11463 (1999).

12. Elson, G. C. et al. J. Immunol. 161,1371–1379 (1998).

13. Davis, S. et al. Science 259, 1736–1739(1993).

14. Gately, M. K. et al. Annu. Rev. Immun. 16,495–521 (1998).

15. Alexander, W. S. et al. Curr. Biol. 9, 605–608(1999).

ously, Elson and colleagues cloned anovel cytokine receptor with homologyto type-I cytokine receptors, which theytermed cytokine-like factor-1 (CLF)12.Intriguingly, they found that CLF was asecreted protein, and therefore predict-ed that it might act more like a ligandthan a receptor. They proposed a func-tional analogy to CNTFRα , which isknown (upon its release from the plas-ma membrane by phospholipase cleav-age of its GPI linkage) to confer CNTFresponsiveness to cells that express onlythe two transmembrane components ofthe CNTF receptor. The ‘ligand’ in thiscase was proposed to be formed by acomplex of CNTF together with the‘soluble’ form of CNTFRα13.

Based on this idea of a secretedreceptor bound to its ligand, Elson andcolleagues demonstrate that a complexof CLC and CLF forms a ligand for theCNTF receptor. First, they show thatCLC is normally not released from cells,but that coexpression of the receptor-like CLF efficiently shepherds CLCthrough the secretion process. Theythen show that CLC and CLF are secret-ed as a stable complex, and that this het-eromer binds to CNTFRα . Critically,each of the subunits of the CNTF receptorcomplex—gp130, LIFRβ and CNTFRα—is required to support CLC–CLF bind-ing to the receptor complex and toactivate downstream signal transductionevents, including phosphorylation ofSTAT3. Finally, and importantly, theyshow that CLC–CLF competes againstCNTF for binding to the CNTF recep-tor and can promote the survival ofembryonic motor neurons in vitro, ahallmark of the biological activity ofCNTF.

The final ligand–receptor complex,containing the CLC–CLF heteromericligand and the CNTFRα , gp130 andLIFR receptor complex, bears strikingsimilarity to the interleukin-12 factorand receptor complex, a configurationthat previously had no known parallelin a neurobiological context. In the caseof interleukin-12, a hematopoetic fac-tor involved in the differentiation andregulation of T-helper cells, the p35 andp40 components of the ligand complexalso require coexpression for secretion,and are further stabilized as a het-erodimer by a disulfide bond14 (Fig. 1).

These elegant experiments show thatCLC–CLF is an excellent candidate forthe elusive CNTF II. Indeed, an earlierstudy has already shown that geneticdisruption of CLF (independently iso-

© 2000 Nature America Inc. • http://neurosci.nature.com©

200

0 N

atu

re A

mer

ica

Inc.

• h

ttp

://n

euro

sci.n

atu

re.c

om