- Email: [email protected]
A Smell to Die for
Developmental Cell
Previews described above. Given the previously determined direct associations of EB1 with dynactin (Ligon et al., 2003), dynactin with cytoplasmic dynein, and dynein with b-catenin, a composite mechanism can be proposed for these interactions (Figure 1). Aspects of this model require further investigation. For example, EB1 is generally thought to be associated specifically with dynamically growing microtubule plus ends, and not to associate strongly with paused microtubules, the presumed state of microtubules tethered at the cortex. Also, this work was examined in a HeLa cell model system, so the relevance to the dynamics of gap junction assembly in vivo must be determined. It will also be of interest to examine if defects in this dynamic assembly correlate with dysfunction or disease. While further studies will be required to fully explore the mechanisms involved, this new work by Shaw et al. (2007) provides a clear functional role for the targeted delivery of cortical components in a polarized manner. Remarkably, aspects of this mechanism appear to be conserved in a number of diverse systems. Recent work on the immune synapse has shown that many of the same proteins are involved in linking microtubules to the cortex to facilitate targeted delivery to sites of cell contact. The immune synapse is the structure that forms when
T cells bind to antigen-presenting cells. The MTOC of the T cell translocates toward the site of cell-cell contact. This translocation is mediated by microtubules projecting from the MTOC toward a ring at the synapse formed in part by the scaffolding protein ADAP. The polarized microtubule cytoskeleton then allows for the directed delivery of lytic granules to the target cell (Stinchcombe et al., 2006). Recent work from Combs et al. (2006) has shown that both dynein and b-catenin localize to the ADAP ring. Loss of ADAP prevents recruitment of dynein to the synapse, and also prevents the reorientation of the MTOC, suggesting that the localization of dynein to the cortex is required to tether microtubules and to reel in the MTOC. There is also tantalizing evidence that the same pathways are involved in other processes, such as cell migration, cytokinesis, and asymmetric cell division. The conserved nature of the core mechanism described here suggests that the cortical capture of microtubules by dynein, dynactin, and the microtubule tip tracking protein EB1 may serve as a powerful mechanism to direct preferential targeting within the cell. This mechanism allows the cell to more efficiently traffic materials, allowing for the dynamic assembly of gap junctions as described in the current work (Shaw et al., 2007), the efficient assembly of adhesion
junctions (Chen et al., 2003; Yanagisawa et al., 2004), or the polarized secretion of lytic granules (Combs et al., 2006; Stinchcombe et al., 2006). Further examples of this type are likely to be identified as the dynamics of the cell are explored in more detail. REFERENCES Burkhardt, J.K., McIlvain, J.M., Jr., Sheetz, M.P., and Argon, Y. (1993). J. Cell Sci. 104, 151–162. Chen, X., Kojima, S., Borisy, G.G., and Green, K.J. (2003). J. Cell Biol. 163, 547–557. Combs, J., Kim, S.J., Tan, S., Ligon, L.A., Holzbaur, E.L., Kuhn, J., and Poenie, M. (2006). Proc. Natl. Acad. Sci. USA 103, 14883–14888. Lauf, U., Giepmans, B.N., Lopez, P., Braconnot, S., Chen, S.C., and Falk, M.M. (2002). Proc. Natl. Acad. Sci. USA 99, 10446– 10451. Ligon, L.A., Karki, S., Tokito, M., and Holzbaur, E.L. (2001). Nat. Cell Biol. 3, 913–917. Ligon, L.A., Shelly, S.S., Tokito, M., and Holzbaur, E.L. (2003). Mol. Biol. Cell 14, 1405–1417. Shaw, R.M., Fay, A.J., Puthenveedu, M.A., von Zastrow, M., Jan, Y.-N., and Jan, L.Y. (2007). Cell 128, 547–560. Stinchcombe, J.C., Majorovits, E., Bossi, G., Fuller, S., and Griffiths, G.M. (2006). Nature 443, 462–465. Yanagisawa, M., Kaverina, I.N., Wang, A., Fujita, Y., Reynolds, A.B., and Anastasiadis, P.Z. (2004). J. Biol. Chem. 279, 9512–9521.
A Smell to Die for Marc Tatar1,* 1
Department Ecology and Evolutionary Biology, Brown University, Providence, RI 02912, USA *Correspondence: [email protected] DOI 10.1016/j.devcel.2007.02.013
How dietary restriction extends life span may involve anticipatory neurons. The odor of dietary yeast is sufficient to speed Drosophila aging while mutants of an odorant-binding protein extend life span. If the yeasty aroma of freshly baked bread awakens your hunger, what might this sort of smell do to the fly Drosophila melanogaster? In a recent report by Lib-
ert, Pletcher and colleagues (Libert et al., 2007), it would seem that the very odor of yeast is sufficient to accelerate aging. This remarkable finding offers opportu-
322 Developmental Cell 12, March 2007 ª2007 Elsevier Inc.
nities to unravel how dietary restriction modulates animal senescence and may help us to better understand our own relationship between nutrition and aging.
Developmental Cell
Previews Drosophilids are fungal feeding flies. Many species specialize on varieties of mushroom, while others such as D. melanogaster thrive on yeasts that grow in decaying fruit. Nutrients from these fungi are required for larval growth. In adults, yeast is consumed to make eggs, and females produce few eggs when yeast is limited. Yeast restriction of adults also extends life span. Indeed, dietary restriction (DR) retards aging in many animals, but how it does so is essentially unknown (Mobbs et al., 2007). In the context of Drosophila it is often thought that DR somehow shifts the allocation of limited nutrients from reproduction toward survivalensuring somatic maintenance. The work of Libert et al. (2007) demonstrates that the mechanisms are not so simple. Odorant binding proteins are expressed in sensory neurons of Drosophila. Previous data showed that DR induced many of these binding proteins. To address whether yeastrelated odors play a direct role in the control of aging, Libert et al. exposed adults to yeast odorant when flies were maintained on yeast-rich or restricted diets. As usual, adults upon restricted diet were long-lived relative to flies upon rich media, but life span reverted toward that of full fed animals when restricted flies were also exposed to yeast odorant. Since odorant exposure did not affect life span of flies upon rich diet, Libert et al. concluded that the survival response to yeast odorant is a regulated biological process within healthy animals (Libert et al., 2007). Libert et al. (2007) excluded a potentially trivial explanation whereby odor exposure increases feeding rates upon restricted diet. They also ruled out an explanation based on reproduction since restricted females with and without yeast odorant produced the same number of eggs. This observation is consistent with accumulating evidence that reproduction and life span can be physiologically decoupled in D. melanogaster, suggesting that the mechanism of dietary restriction must involve a mechanism beyond the simple idea of resource allocation (Leroi, 2001).
If yeast odor is sufficient to reduce life span of dietary restricted adults, could inhibition of odor sensing be sufficient to extend longevity in full fed flies? To address this question Libert et al. tested mutants of Or83b, an odorant binding protein that is broadly expressed in sensory neurons. Loss of Or83b blocked behavioral responses to yeast odor and impressively increased life span. Similar extension of longevity was produced by disruption of G protein signals in sensory neurons. Together these results suggest there is a specific system of sensory signaling with the capacity to affect broad control of survival assurance physiology. From these data it would seem there is a common mechanism underlying the longevity benefit conferred by blocking odor sensing and by DR. To test this idea Libert et al. measured lifespan of Or83b adults maintained on a range of diets. Wild-type females increased adjusted life expectancy up to 25% on restricted diets relative to rich diets. Mutants of Or83b also responded to DR, but with responses ranging only from 6% to 11%. Thus, Or83b is not essential for DR to increase life span but its loss diminishes the efficiency of DR to affect longevity. Libert et al. concluded that the mechanism whereby Or83b increases life span is largely diet independent. Others may interpret this result differently: odorant sensing affected by a single gene accounts for at least half of improved longevity in response to diet, and thus may play a major role in how DR modulates aging. There is no question that some type of systemic signaling must mediate how odor sensing affects aging. Only a few neurons coordinate the whole body to ensure longevity and, as also demonstrated, stress resistance. Gustatory and sensory neurons can regulate lifespan of C. elegans (Alcedo and Kenyon, 2004; Apfeld and Kenyon, 1999). Although the cues are unknown, the insulin/IGF responsive DAF-16/FOXO transcription factor is required to some extent for these mutants to increase worm life span. To investigate a role of insulin signals in odor sensory control of Drosophila aging, Libert et al. (2007) measured mes-
sage of the seven insulin-like peptides (ilp). Only ilp3 decreased in Or83b mutants, in contrast to reduction of ilp2 reported elsewhere upon manipulation of dFOXO or JNK (Hwangbo et al., 2004; Wang et al., 2005). Message of the insulin-regulated gene thor was measured to assess the potential systemic consequences of reduced ilp3. Rather than increase as expected if insulin signaling were repressed, thor transcripts were reduced. From these few measures there is no compelling association between Or83b and insulin to implicate this hormone as the mediator of aging by odor sensing. It shall be important to investigate alternative hormone systems of insects that respond to nutrition, such as juvenile hormone, ecdysone, and adipokinetic hormone (an analog of mammalian glucagon). At the same time, further data are needed to exclude insulin signaling since nutrient control of Drosophila insulin-like peptides involve posttranslational processes (Tu and Tatar, 2003), and the in vivo affects of FOXO upon targets remains poorly defined (Gershman et al., 2006). Testing how Or83b affects life span when crossed to FOXO null mutant should be a priority. From the view of humans we might take away a mixed message from these results. Simply passing through a bakery may now be added to the list of longevity risk factors—but perhaps only for the few of us that are not consuming the equivalent of a full diet (and only if yeast were our primary nutrient). On perhaps the more serious side, these results imply there are high order sensory and hormonal systems within animals that modulate longevity assurance using cues associated with anticipated nutrient state. In principle we might someday identify how these anticipatory systems work and find ways to manipulate their analogs within us to improve our health during the process of normal aging.
REFERENCES Alcedo, J., and Kenyon, C. (2004). Neuron 41, 45–55. Apfeld, J., and Kenyon, C. (1999). Nature 402, 804–809.
Developmental Cell 12, March 2007 ª2007 Elsevier Inc. 323
Developmental Cell
Previews Gershman, B., Puig, O., Hang, L., Peitzsh, R.M., Tatar, M., and Garofalo, R.S. (2006). Physiol. Genomics, in press. Published online November 7, 2006. 10.1152/physiolgenomics. 00061.2006. Hwangbo, D.S., Gershman, B., Tu, M.P., Palmer, M.R., and Tatar, M. (2004). Nature 429, 562–566.
Leroi, A.M. (2001). Trends Ecol. Evol. 16, 24–29.
Mobbs, C.V., Yen, K., and Hof, P.R. (2007). Interdisciplinary Topics in Gerontology, Vol 35 (Basel: Karger).
Libert, S., Zwiener, J., Chu, X., VanVoorhies, W., Roman, G., and Pletcher, S.D. (2007). Science, in press. Published online February 1, 2007. 10.1126/science. 1136610.
Tu, M.P., and Tatar, M. (2003). Aging Cell 2, 327–333. Wang, M.C., Bohmann, D., and Jasper, H. (2005). Cell 121, 115–125.
‘‘Smad’’eningly Erratic: Target Gene Methylation Determines Whether TGFb Promotes or Suppresses Malignant Glioma Santosh Kesari,1,2 Laurie Jackson-Grusby,3 and Charles D. Stiles1,* 1
Dana-Farber Cancer Institute, Boston, Massachusetts 02115, USA Department of Neurology, Brigham and Women’s Hospital, Boston, Massachusetts 02115, USA 3 Children’s Hospital Boston, Boston, Massachusetts 02115, USA *Correspondence: [email protected] DOI 10.1016/j.devcel.2007.02.008 2
TGFb functions as a tumor suppressor in some contexts and a tumor promoter in others. In a recent issue of Cancer Cell, Bruna et al. (2007) shed light on an epigenetic mechanism that underlies this schizophrenic behavior in malignant glioma. Their findings highlight a stem cell/cancer link.and a potential blind spot in large-scale cancer genome sequencing projects. Transforming growth factor beta (TGFb) is the founding member and namesake of a large family of cytokines characterized by a set of conserved cysteine residues that fold the proteins into rigid ‘‘cysteine knot’’ motifs through formation of intrachain disulfide bridges. TGFb proteins activate heterodimeric cell surface receptors that function as serine/threonine protein kinases. Signaling information is relayed from these cell surface receptors to the cell nucleus via phosphorylation of SMAD proteins. Phosphorylated SMADs relocalize from the inner cell surface to the nucleus where they function as transcription factors. TGFbs play multifaceted roles in embryonic development and tissue homeostasis (Massague and Gomis, 2006). However they have especially prominent functions in the area of cell cycle control. In cell culture, nanomolar concentrations of TGFb suppress the growth of normal epithelial cells and lymphocytes. In transgenic mouse models,
misexpression of TGFb retards the formation of mammary ductal epithelium and delays the onset of mammary carcinomas induced by chemical and viral carcinogens (see Bierie and Moses [2006] and references therein). Some human cancers acquire lossof-function mutations in the TGFb signaling pathway that would ablate these cytostatic responses (Chen et al., 2001). These observations and other data are consistent with the view that TGFb and its downstream signal generators function as classical tumor suppressors . but there is another side to the story. In mice with previously established mammary carcinomas, TGFb actually promotes tumor invasion and metastasis (Bierie and Moses [2006] and references therein). In murine models of skin carcinoma, TGFb functions as a tumor suppressor at early stages of tumor development and as a tumor accelerant at later stages. Some human tumors retain an intact TGFb signaling apparatus while avoiding
324 Developmental Cell 12, March 2007 ª2007 Elsevier Inc.
growth arrest. In malignant gliomas, things go even further than that. In a recent issue of Cancer Cell, Bruna and colleagues build upon mounting evidence that TGFb has proliferative/oncogenic functions in gliomas that are channeled through PDGF autocrine/paracrine loops (Bruna et al. [2007] and references therein). The point of departure for their study is the observation that activation of the TGFb signaling apparatus serves as a negative prognostic indicator for glioma patients. A direct correlation with proliferation index and grade of tumor and an inverse correlation with survival provide compelling evidence that activation of the TGFb signaling pathway provides a selective advantage to tumor cells. How can a potent antimitotic agent work to the advantage of malignant gliomas? Some of the late-stage tumorpromoting actions of TGFb in murine models (i.e., enhanced cell motility and invasion, induction of angiogenesis, immune suppression) appear to
Previews described above. Given the previously determined direct associations of EB1 with dynactin (Ligon et al., 2003), dynactin with cytoplasmic dynein, and dynein with b-catenin, a composite mechanism can be proposed for these interactions (Figure 1). Aspects of this model require further investigation. For example, EB1 is generally thought to be associated specifically with dynamically growing microtubule plus ends, and not to associate strongly with paused microtubules, the presumed state of microtubules tethered at the cortex. Also, this work was examined in a HeLa cell model system, so the relevance to the dynamics of gap junction assembly in vivo must be determined. It will also be of interest to examine if defects in this dynamic assembly correlate with dysfunction or disease. While further studies will be required to fully explore the mechanisms involved, this new work by Shaw et al. (2007) provides a clear functional role for the targeted delivery of cortical components in a polarized manner. Remarkably, aspects of this mechanism appear to be conserved in a number of diverse systems. Recent work on the immune synapse has shown that many of the same proteins are involved in linking microtubules to the cortex to facilitate targeted delivery to sites of cell contact. The immune synapse is the structure that forms when
T cells bind to antigen-presenting cells. The MTOC of the T cell translocates toward the site of cell-cell contact. This translocation is mediated by microtubules projecting from the MTOC toward a ring at the synapse formed in part by the scaffolding protein ADAP. The polarized microtubule cytoskeleton then allows for the directed delivery of lytic granules to the target cell (Stinchcombe et al., 2006). Recent work from Combs et al. (2006) has shown that both dynein and b-catenin localize to the ADAP ring. Loss of ADAP prevents recruitment of dynein to the synapse, and also prevents the reorientation of the MTOC, suggesting that the localization of dynein to the cortex is required to tether microtubules and to reel in the MTOC. There is also tantalizing evidence that the same pathways are involved in other processes, such as cell migration, cytokinesis, and asymmetric cell division. The conserved nature of the core mechanism described here suggests that the cortical capture of microtubules by dynein, dynactin, and the microtubule tip tracking protein EB1 may serve as a powerful mechanism to direct preferential targeting within the cell. This mechanism allows the cell to more efficiently traffic materials, allowing for the dynamic assembly of gap junctions as described in the current work (Shaw et al., 2007), the efficient assembly of adhesion
junctions (Chen et al., 2003; Yanagisawa et al., 2004), or the polarized secretion of lytic granules (Combs et al., 2006; Stinchcombe et al., 2006). Further examples of this type are likely to be identified as the dynamics of the cell are explored in more detail. REFERENCES Burkhardt, J.K., McIlvain, J.M., Jr., Sheetz, M.P., and Argon, Y. (1993). J. Cell Sci. 104, 151–162. Chen, X., Kojima, S., Borisy, G.G., and Green, K.J. (2003). J. Cell Biol. 163, 547–557. Combs, J., Kim, S.J., Tan, S., Ligon, L.A., Holzbaur, E.L., Kuhn, J., and Poenie, M. (2006). Proc. Natl. Acad. Sci. USA 103, 14883–14888. Lauf, U., Giepmans, B.N., Lopez, P., Braconnot, S., Chen, S.C., and Falk, M.M. (2002). Proc. Natl. Acad. Sci. USA 99, 10446– 10451. Ligon, L.A., Karki, S., Tokito, M., and Holzbaur, E.L. (2001). Nat. Cell Biol. 3, 913–917. Ligon, L.A., Shelly, S.S., Tokito, M., and Holzbaur, E.L. (2003). Mol. Biol. Cell 14, 1405–1417. Shaw, R.M., Fay, A.J., Puthenveedu, M.A., von Zastrow, M., Jan, Y.-N., and Jan, L.Y. (2007). Cell 128, 547–560. Stinchcombe, J.C., Majorovits, E., Bossi, G., Fuller, S., and Griffiths, G.M. (2006). Nature 443, 462–465. Yanagisawa, M., Kaverina, I.N., Wang, A., Fujita, Y., Reynolds, A.B., and Anastasiadis, P.Z. (2004). J. Biol. Chem. 279, 9512–9521.
A Smell to Die for Marc Tatar1,* 1
Department Ecology and Evolutionary Biology, Brown University, Providence, RI 02912, USA *Correspondence: [email protected] DOI 10.1016/j.devcel.2007.02.013
How dietary restriction extends life span may involve anticipatory neurons. The odor of dietary yeast is sufficient to speed Drosophila aging while mutants of an odorant-binding protein extend life span. If the yeasty aroma of freshly baked bread awakens your hunger, what might this sort of smell do to the fly Drosophila melanogaster? In a recent report by Lib-
ert, Pletcher and colleagues (Libert et al., 2007), it would seem that the very odor of yeast is sufficient to accelerate aging. This remarkable finding offers opportu-
322 Developmental Cell 12, March 2007 ª2007 Elsevier Inc.
nities to unravel how dietary restriction modulates animal senescence and may help us to better understand our own relationship between nutrition and aging.
Developmental Cell
Previews Drosophilids are fungal feeding flies. Many species specialize on varieties of mushroom, while others such as D. melanogaster thrive on yeasts that grow in decaying fruit. Nutrients from these fungi are required for larval growth. In adults, yeast is consumed to make eggs, and females produce few eggs when yeast is limited. Yeast restriction of adults also extends life span. Indeed, dietary restriction (DR) retards aging in many animals, but how it does so is essentially unknown (Mobbs et al., 2007). In the context of Drosophila it is often thought that DR somehow shifts the allocation of limited nutrients from reproduction toward survivalensuring somatic maintenance. The work of Libert et al. (2007) demonstrates that the mechanisms are not so simple. Odorant binding proteins are expressed in sensory neurons of Drosophila. Previous data showed that DR induced many of these binding proteins. To address whether yeastrelated odors play a direct role in the control of aging, Libert et al. exposed adults to yeast odorant when flies were maintained on yeast-rich or restricted diets. As usual, adults upon restricted diet were long-lived relative to flies upon rich media, but life span reverted toward that of full fed animals when restricted flies were also exposed to yeast odorant. Since odorant exposure did not affect life span of flies upon rich diet, Libert et al. concluded that the survival response to yeast odorant is a regulated biological process within healthy animals (Libert et al., 2007). Libert et al. (2007) excluded a potentially trivial explanation whereby odor exposure increases feeding rates upon restricted diet. They also ruled out an explanation based on reproduction since restricted females with and without yeast odorant produced the same number of eggs. This observation is consistent with accumulating evidence that reproduction and life span can be physiologically decoupled in D. melanogaster, suggesting that the mechanism of dietary restriction must involve a mechanism beyond the simple idea of resource allocation (Leroi, 2001).
If yeast odor is sufficient to reduce life span of dietary restricted adults, could inhibition of odor sensing be sufficient to extend longevity in full fed flies? To address this question Libert et al. tested mutants of Or83b, an odorant binding protein that is broadly expressed in sensory neurons. Loss of Or83b blocked behavioral responses to yeast odor and impressively increased life span. Similar extension of longevity was produced by disruption of G protein signals in sensory neurons. Together these results suggest there is a specific system of sensory signaling with the capacity to affect broad control of survival assurance physiology. From these data it would seem there is a common mechanism underlying the longevity benefit conferred by blocking odor sensing and by DR. To test this idea Libert et al. measured lifespan of Or83b adults maintained on a range of diets. Wild-type females increased adjusted life expectancy up to 25% on restricted diets relative to rich diets. Mutants of Or83b also responded to DR, but with responses ranging only from 6% to 11%. Thus, Or83b is not essential for DR to increase life span but its loss diminishes the efficiency of DR to affect longevity. Libert et al. concluded that the mechanism whereby Or83b increases life span is largely diet independent. Others may interpret this result differently: odorant sensing affected by a single gene accounts for at least half of improved longevity in response to diet, and thus may play a major role in how DR modulates aging. There is no question that some type of systemic signaling must mediate how odor sensing affects aging. Only a few neurons coordinate the whole body to ensure longevity and, as also demonstrated, stress resistance. Gustatory and sensory neurons can regulate lifespan of C. elegans (Alcedo and Kenyon, 2004; Apfeld and Kenyon, 1999). Although the cues are unknown, the insulin/IGF responsive DAF-16/FOXO transcription factor is required to some extent for these mutants to increase worm life span. To investigate a role of insulin signals in odor sensory control of Drosophila aging, Libert et al. (2007) measured mes-
sage of the seven insulin-like peptides (ilp). Only ilp3 decreased in Or83b mutants, in contrast to reduction of ilp2 reported elsewhere upon manipulation of dFOXO or JNK (Hwangbo et al., 2004; Wang et al., 2005). Message of the insulin-regulated gene thor was measured to assess the potential systemic consequences of reduced ilp3. Rather than increase as expected if insulin signaling were repressed, thor transcripts were reduced. From these few measures there is no compelling association between Or83b and insulin to implicate this hormone as the mediator of aging by odor sensing. It shall be important to investigate alternative hormone systems of insects that respond to nutrition, such as juvenile hormone, ecdysone, and adipokinetic hormone (an analog of mammalian glucagon). At the same time, further data are needed to exclude insulin signaling since nutrient control of Drosophila insulin-like peptides involve posttranslational processes (Tu and Tatar, 2003), and the in vivo affects of FOXO upon targets remains poorly defined (Gershman et al., 2006). Testing how Or83b affects life span when crossed to FOXO null mutant should be a priority. From the view of humans we might take away a mixed message from these results. Simply passing through a bakery may now be added to the list of longevity risk factors—but perhaps only for the few of us that are not consuming the equivalent of a full diet (and only if yeast were our primary nutrient). On perhaps the more serious side, these results imply there are high order sensory and hormonal systems within animals that modulate longevity assurance using cues associated with anticipated nutrient state. In principle we might someday identify how these anticipatory systems work and find ways to manipulate their analogs within us to improve our health during the process of normal aging.
REFERENCES Alcedo, J., and Kenyon, C. (2004). Neuron 41, 45–55. Apfeld, J., and Kenyon, C. (1999). Nature 402, 804–809.
Developmental Cell 12, March 2007 ª2007 Elsevier Inc. 323
Developmental Cell
Previews Gershman, B., Puig, O., Hang, L., Peitzsh, R.M., Tatar, M., and Garofalo, R.S. (2006). Physiol. Genomics, in press. Published online November 7, 2006. 10.1152/physiolgenomics. 00061.2006. Hwangbo, D.S., Gershman, B., Tu, M.P., Palmer, M.R., and Tatar, M. (2004). Nature 429, 562–566.
Leroi, A.M. (2001). Trends Ecol. Evol. 16, 24–29.
Mobbs, C.V., Yen, K., and Hof, P.R. (2007). Interdisciplinary Topics in Gerontology, Vol 35 (Basel: Karger).
Libert, S., Zwiener, J., Chu, X., VanVoorhies, W., Roman, G., and Pletcher, S.D. (2007). Science, in press. Published online February 1, 2007. 10.1126/science. 1136610.
Tu, M.P., and Tatar, M. (2003). Aging Cell 2, 327–333. Wang, M.C., Bohmann, D., and Jasper, H. (2005). Cell 121, 115–125.
‘‘Smad’’eningly Erratic: Target Gene Methylation Determines Whether TGFb Promotes or Suppresses Malignant Glioma Santosh Kesari,1,2 Laurie Jackson-Grusby,3 and Charles D. Stiles1,* 1
Dana-Farber Cancer Institute, Boston, Massachusetts 02115, USA Department of Neurology, Brigham and Women’s Hospital, Boston, Massachusetts 02115, USA 3 Children’s Hospital Boston, Boston, Massachusetts 02115, USA *Correspondence: [email protected] DOI 10.1016/j.devcel.2007.02.008 2
TGFb functions as a tumor suppressor in some contexts and a tumor promoter in others. In a recent issue of Cancer Cell, Bruna et al. (2007) shed light on an epigenetic mechanism that underlies this schizophrenic behavior in malignant glioma. Their findings highlight a stem cell/cancer link.and a potential blind spot in large-scale cancer genome sequencing projects. Transforming growth factor beta (TGFb) is the founding member and namesake of a large family of cytokines characterized by a set of conserved cysteine residues that fold the proteins into rigid ‘‘cysteine knot’’ motifs through formation of intrachain disulfide bridges. TGFb proteins activate heterodimeric cell surface receptors that function as serine/threonine protein kinases. Signaling information is relayed from these cell surface receptors to the cell nucleus via phosphorylation of SMAD proteins. Phosphorylated SMADs relocalize from the inner cell surface to the nucleus where they function as transcription factors. TGFbs play multifaceted roles in embryonic development and tissue homeostasis (Massague and Gomis, 2006). However they have especially prominent functions in the area of cell cycle control. In cell culture, nanomolar concentrations of TGFb suppress the growth of normal epithelial cells and lymphocytes. In transgenic mouse models,
misexpression of TGFb retards the formation of mammary ductal epithelium and delays the onset of mammary carcinomas induced by chemical and viral carcinogens (see Bierie and Moses [2006] and references therein). Some human cancers acquire lossof-function mutations in the TGFb signaling pathway that would ablate these cytostatic responses (Chen et al., 2001). These observations and other data are consistent with the view that TGFb and its downstream signal generators function as classical tumor suppressors . but there is another side to the story. In mice with previously established mammary carcinomas, TGFb actually promotes tumor invasion and metastasis (Bierie and Moses [2006] and references therein). In murine models of skin carcinoma, TGFb functions as a tumor suppressor at early stages of tumor development and as a tumor accelerant at later stages. Some human tumors retain an intact TGFb signaling apparatus while avoiding
324 Developmental Cell 12, March 2007 ª2007 Elsevier Inc.
growth arrest. In malignant gliomas, things go even further than that. In a recent issue of Cancer Cell, Bruna and colleagues build upon mounting evidence that TGFb has proliferative/oncogenic functions in gliomas that are channeled through PDGF autocrine/paracrine loops (Bruna et al. [2007] and references therein). The point of departure for their study is the observation that activation of the TGFb signaling apparatus serves as a negative prognostic indicator for glioma patients. A direct correlation with proliferation index and grade of tumor and an inverse correlation with survival provide compelling evidence that activation of the TGFb signaling pathway provides a selective advantage to tumor cells. How can a potent antimitotic agent work to the advantage of malignant gliomas? Some of the late-stage tumorpromoting actions of TGFb in murine models (i.e., enhanced cell motility and invasion, induction of angiogenesis, immune suppression) appear to