Id and Development

Id and development

Yoshifumi Yokota*,1

1Department of Biochemistry, Fukui Medical University, Shimoaizuki 23-3, Matsuoka, Fukui 910-1193, Japan

During development, it is obvious that enormous multi- variety of tissues and cell types (Olson and Klein, plication and diversi®cation of cells is required to build a 1994; Lee, 1997; Massari and Murre, 2000). In general, body plan from a single fertilized egg and that these two heterodimers formed between tissue-speci®c and ubi- processes, proliferation and dierentiation, must be quitous bHLH factors function as active transcription coordinated properly. Id proteins, negative regulators of factors and promote cell dierentiation by activating basic helix ± loop ± helix transcription factors, possess the tissue-speci®c gene expression. On the other hand, Id ability to inhibit dierentiation and to stimulate proteins, lacking the basic domain required for DNA- proliferation, and are useful molecules for investigating binding, dimerize with and quench primarily the the mechanisms regulating development. In the past few ubiquitous bHLH factors (Benezra et al., 1990; Jen et years, our understanding of the roles of Id proteins has al., 1992; Norton et al., 1998) and thereby inhibit cell been substantially enhanced by the detailed investigation dierentiation (Jen et al., 1992; Norton et al., 1998). It of genetically modi®ed animals. The data have indicated is clear that inhibitory and stimulatory factors are that the functions of Id proteins in vivo are functionally important in achieving a correct body plan and in this related to those revealed by earlier work in cell culture sense, the Id proteins are beginning to be shown to systems. However, unexpected organs and cell types have play a central role. also been found to require Id proteins for their normal Gene-targeted mice have been successfully generated development. This review looks at the advances made in for each Id family member and studies of three of them our understanding of the in vivo functions of Id proteins. (for Id1, Id2 and Id3) have been reported thus far (Yan The topics discussed include neurogenesis, natural killer et al., 1997; Yokota et al., 1999; Lyden et al., 1999, cell development, lymphoid organogenesis, mammary Rivera et al., 2000). Overexpression studies of Id gland development and spermatogenesis. Oncogene proteins in several in vivo systems have also been (2001) 20, 8290 ± 8298 carried out (Sun, 1994; Martinsen and Bronner-Fraser, 1998; Wice and Gordon, 1998; Morrow et al., 1999; Keywords: Id; basic helix ± loop ± helix (bHLH); dier- Kim et al., 1999; Cai et al., 2000). Phenotypes detected entiation; proliferation; development in both gain-of-function and loss-of-function experiments give us insight into potential bHLH proteins as well as regulatory roles of Id proteins in the respective Introduction organ development. In the last couple of years, great advances have been made in identifying the roles of Id During development, cells proliferate rapidly to build proteins based on studies of the genetically modi®ed our bodies and dierentiate into the various cell types animals and culture systems. The ®ndings are relevant that organize a sophisticated living system. To carry not only for people working with Id proteins but also out and support normal development, cell proliferation more broadly for those working in developmental and and dierentiation are under the control of orches- cancer biology. This review will summarize recent trated, extracellular regulatory signals and of a progress in Id biology by focusing on the results genetically de®ned intrinsic program. Id proteins, obtained with Id-de®cient mice and discuss how inhibitors of basic helix ± loop ± helix (bHLH) factors, dierentiation and proliferation are intricately regu- are positive and negative regulators of proliferation lated by Id proteins in vivo. and dierentiation, respectively, and provide useful tools for investigating the mechanisms underlying cell proliferation and dierentiation controls (Norton et al., Early development 1998). Four members, Id1 through Id4, have been identi®ed in mammals (Benezra et al., 1990; Sun et al., During development, all members of the Id family are 1991; Christy et al., 1991; Riechmann et al., 1994). expressed in many organs and tissues, showing over- It is well known that many bHLH transcription lapping but distinct expression patterns in develop- factors positively regulate cell dierentiation in a mental stage-dependent and region-speci®c manners (Evans and O'Brien, 1993; Riechmann and Sablitzky, 1995; Zhu et al., 1995; Jen et al., 1996, 1997). This *Correspondence: Y Yokota; re¯ects the important roles of Id proteins in the E-mail: [email protected] regulation of bHLH factors' functions in normal Id and development Y Yokota 8291 development and dierentiation. As demonstrated by least in neurons in layers II/III and V of the adult in situ hybridization studies, a general feature of Id cerebral cortex. The proliferative potential of these Id gene expression during development is that the proteins may be restrained by the action of the pRB distributions of Id1, Id2 and Id3 are overlapping, tumor suppressor protein as discussed by Lasorella et while Id4 shows a unique pattern during embryogenesis al. in this issue. Compared to other members that are (Riechmann and Sablitzky, 1995; Jen et al., 1996, also expressed in many peripheral tissues, Id4 is rather 1997). In addition, tissues that undergo epithelial- selectively expressed in the nervous system during mesenchymal interactions have a tendency to express embryogenesis (Riechmann and Sablitzky, 1995; Jen Id genes, as observed in the salivary gland, kidney and et al., 1997). lung (Evans and O'Brien, 1993; Zhu et al., 1995; Jen et No apparent phenotype is detected in the central al., 1996). However, inactivation of a single Id gene nervous system of mice lacking either Id1 or Id3 seems not to be sucient to cause a defect in early (Lyden et al., 1999). In Id27/7 mice, the olfactory mouse development. bulb is small and the frontal cortex is slightly narrow; Id2 has been shown to play a role in cell fate the mechanisms of these eects are now under speci®cation of neural crest cells at least in the chick investigation (Yokota, unpublished observations). His- (Martinsen and Bronner-Fraser, 1998). At the 5-somite tological analyses of the Id27/7 brain using routine stage of the chick embryo, Id2 is expressed in the most staining methods have failed to identify any apparent dorsal cranial neural fold region of the presumptive anomaly so far, including any in the layer formation of brain. At the 11-somite stage, its expression is found in the cerebral cortex or the morphology of Purkinje cells the neural tube and in a subpopulation of migrating in the cerebellum (Kitajima and Yokota, unpublished cranial neural crest cells at the midbrain level. Retro- observations). The subtle eects of the inactivation of virus-mediated overexpression of Id2 on the embryonic the single Id genes may be the result of compensation surface induces the disappearance of the ectoderm via the functional redundancy of other Id proteins, overlying the neural tube and conversion of these cells because their expression domains overlap extensively. to neural crest cells. Id2-overexpression also results in In support of this notion, when Id1 and Id3 genes are overgrowth of the dorsal neural tube and the inactivated simultaneously, the compound mutant mice formation of ganglion-like aggregates within the die in utero and have small brains (Lyden et al., 1999). cranial neural tube, depending on the timing of The embryonic lethality is ascribable to a defect in retrovirus infection. In contrast, overexpression of Id2 angiogenesis (reviewed by Benezra and Ra®i in this in the underlying mesenchyme does not cause such issue). In the Id17/7Id37/7 embryonic brain, no eects. These observations suggest that a bHLH factor signi®cant increase in the number of apoptotic cells is bearing an epidermalizing activity exists in the found but premature withdrawal of neuroblasts from ectoderm and the functional regulation of the factor the cell cycle is evident. Postmitotic neuronal cell by Id2 determines the cell fate. Id2 may also function markers such as microtubule-associated protein 2 are in mammalian embryos in a similar manner, although detected more widely than in normal control mice on Id27/7 mice do not show any obvious phenotypes embryonic day 11.5. This is accompanied by the related to such a function (Yokota, unpublished elevated expression of cyclin-dependent kinase inhibi- observations). tors p16/INK4a and p27/Kip1, and by a decrease in the area positive for a proliferating cell marker, Ki67. In addition, stronger and more widespread expression The nervous system than in controls is observed for both the determination- and dierentiation-related neurogenic bHLH Expression of Id genes is also found in many regions of factors, suggesting that the premature activation of the nervous system in very complex and dynamic the genetic cascade of bHLH factors accelerates patterns. Many groups have reported the detailed neurogenesis. The premature terminal dierentiation spatio-temporal expression patterns of the respective of Id17/7Id37/7 neuroblasts thus appears to result Id genes (Duncan et al., 1992, 1997; Neuman et al., from precocious functioning of neurogenic bHLH 1993; Riechmann and Sablitzky, 1995; Zhu et al., 1995; factors in the absence of their inhibitory regulators. Jen et al., 1996, 1997; Tzeng and de Vellis, 1998; Id47/7 mice were successfully generated and have Andres-Barquin et al., 2000). In general, all of the been analysed by Sablitzky and colleagues. The neural factors are expressed in developing neuroblasts, with phenotypes of these mutant mice are more prominent similar patterns for Id1, Id2 and Id3, and a distinct than those of mice de®cient for any one of the other Id pattern for Id4, as in the developing non-neural tissues. genes (Sablitzky, personal communication). Detailed One remarkable feature of Id gene expression in the histochemical analysis indicates that the brains of neural tissues is that Id2 and Id4 are also expressed in Id47/7 mice are severely malformed. The lateral migrating postmitotic neurons in the intermediate zone ventricle as well as the 3rd and 4th ventricles are and even in speci®c neurons in the adult brain greatly enlarged. On the other hand, the septum and (Neuman et al., 1993; Riechmann and Sablitzky, caudate putamen as well as the thalamus and 1995; Jen et al., 1997; Tzeng and de Vellis, 1998). hypothalamus are reduced in size and/or disorganized. Both Id2 and Id4 are expressed strongly in Purkinje The number of oligodendrocytes is reduced in all of cells in the adult cerebellum and Id2 is expressed at these brain regions. In accordance with this, neural

Oncogene Id and development Y Yokota 8292 stem cells isolated from Id47/7 embryos do not proliferation rate is also found in Id27/7 OPCs. An dierentiate to oligodendrocyte precursors in vivo. interesting ®nding is the relation between oligoden- These results suggest that Id4 function is essential for drocyte dierentiation and the subcellular localization the development of oligodendrocytes. It is, however, of the Id2 protein. In OPCs, Id2 protein is detected in dicult to exclude the possibility that premature the nucleus. During dierentiation induction, however, dierentiation of Id47/7 neural cells leads to Id2 protein moves to the cytoplasm and this transloca- defective cell fate speci®cation in the oligodendrocyte tion precedes the oligodendrocyte dierentiation. The lineage. On the other hand, Cre-mediated overexpres- results obtained with Id27/7 OPCs suggest that Id2 is sion of Id4 preformed by the same group (Sablitzky, not essential for the generation of oligodendrocytes but personal communication) in the developing nervous participates in determining the timing of the dier- system of the mouse stimulates cell proliferation and entiation. Id4 is detected in the nucleus of OPCs but apoptosis in the adult brain. This is consistent with the whether intracellular translocation occurs is unclear observation that Id47/7 neural precursor cells (Kondo and Ra, 2000a). Taken together, the ®ndings proliferate more slowly than their wild-type counter- suggest that it is most probable that the development parts in vivo. In addition, enforced expression of Id4 of oligodendrocytes requires Id4 and the timing of results in a developmental shift from glial to neuronal oligodendrocyte dierentiation from OPCs is deter- cells in the cortex. Thus Id4 appears to be a crucial mined by both Id2 and Id4. While bHLH factors seem regulator of neural stem cell proliferation and fate to regulate the dierentiation process (Kondo and determination, although results obtained by over- Ra, 2000a,b; Wang et al., 2001) it is unclear which are expression studies may include not only Id4-speci®c targeted by the Id proteins. Mash1 is expressed in eects but also those could be caused commonly by all OPCs but it is unlikely to be involved in the process, Id proteins. because oligodendrocyte dierentiation occurs nor- Among the functions of Id4 in the neural develop- mally in the absence of Mash1 (Wang et al., 2001). ment mentioned above, its eects on oligodendrocytes The recently identi®ed bHLH factors olg1/oligo2 and are of great interest. Oligodendrocytes are glial cells olg2/oligo1 are candidates for such regulatory factors that myelinize large axons. When oligodendrocyte (Lu et al., 2000; Zhou et al., 2000). precursor cells (OPCs) isolated from neonatal rat optic nerve are cultured under appropriate conditions, they stop dividing and start to dierentiate after approxi- The immune system mately eight rounds of cell division, suggesting a cell- intrinsic timing mechanism that regulates the multi- Cells in the immune system develop from hematopoie- plicity of cell division as well as a tight link between tic stem cells through multiple processes that restrict proliferation and dierentiation (Temple and Ra, the respective cell fates step by step. Their maturation 1986; Barres et al., 1994; Durand and Ra, 2000). Id4 is under the regulation of sets of several kinds of has been reported to function as the intracellular timer, transcription factors depending on the lineages. Tran- at least in part, of oligodendrocyte dierentiation scription factors with the bHLH motif play pivotal (Kondo and Ra, 2000a). OPCs express all members roles in the development of hematopoietic cells, and Id of the Id gene family and only Id4 shows alterations in proteins are important regulators of these processes. its expression according to the proliferation state of Since the role of the Id proteins in most hematopoietic OPCs. When OPCs are induced to dierentiate in processes is reviewed by Rivera and Murre in this culture by withdrawal of PDGF or addition of thyroid issue, this review will focus on the roles of Id proteins hormone, Id4 expression decreases. Retrovirus- in the development of NK cells and the peripheral mediated enforced expression of Id4 in OPCs results lymphoid organs. in increased proliferation and impaired dierentiation to oligodendrocytes, while Id1 is less eective in this NK cells system. These observations seem to be consistent with the ®ndings obtained from the studies with Id47/7 NK cells belong to a minor population of the lymphoid mice in terms of a role of Id4 in oligodendrocyte lineage cells. However, they are indispensable in development, although similar analysis of OPCs of defence mechanisms such as innate immunity and Id47/7 is not available due to the embryonic lethality tumor rejection. Although little is known about the of Id47/7 mice. On the other hand, Id2 has also been developmental pathway of NK cells compared with demonstrated to participate in oligodendrocyte dier- those of T or B cells, recent studies using gene-targeted entiation (Wang et al., 2001). When OPCs are mice have revealed several genes involved in NK cell transformed with Id2, they exhibit substantially development. One of them is Id2. Id27/7 mice exhibit attenuated ability to dierentiate to oligodendrocytes severely compromised NK cell development, whereas T under the dierentiation-inducing conditions and dis- and B cells develop normally (Yokota et al., 1999; play an accelerated proliferation rate in the presence of Yokota et al., unpublished observation). In Id27/7 PDGF as a mitogen. In cultures of OPCs derived from mice, the number of NK cells is reduced, by 90% Id27/7 mice, premature oligodendrocyte dierentia- relative to wild-type controls in the spleen, bone tion is observed and an increased percentage of marrow, liver, and fetal thymus. The cytolytic activity oligodendrocytes is noted. A decrease in the cell of the Id27/7 splenocytes against tumor cells is

Oncogene Id and development Y Yokota 8293 reduced to a degree comparable to the reduction of the early T cell development. These observations strongly NK cell number, indicating that the NK cells generated suggest that Id2 is critically involved in cell fate in Id27/7 mice seem to be functional. The defect is determination to NK cells. intrinsic to NK precursors and not to the microenvir- Analysis of the regulation of early T cell develop- onment of the bone marrow, because the dierentia- ment may give insight into what happens in bipotent tion of NK cells from Id27/7 bone marrow cells is NK/T precursors to restrict their fate to the NK defective even in the presence of IL-15, which can lineage. The activity of E proteins is required for the substitute for the bone marrow microenvironment normal progression of T cell development. In required for NK cell development (Carson et al., 1994). particular, the transition of early T cells from the NK cells and T cells share common progenitors CD44+CD257 stage to the CD44+CD25+ stage is (Williams et al., 1998; Carlyle and Zuniga-P¯ucker, perturbed in E2A7/7 mice, indicating that E2A 1998; Spits et al., 1998). In the fetal thymus which plays a role in promoting the developmental transition contains bipotent T/NK progenitors, early T cell (Zhuang et al., 1996, 1998; Bain et al., 1997). The development can be divided into four stages according fetal thymocyte fraction at the CD44+CD257 stage to the expression of the cell surface markers CD44 contains bipotent T/NK progenitors, and fetal (pgp-1) and CD25 (interleukin 2 receptor a, IL-2Ra) thymocytes at the CD44+CD25+ stage start to lose (Shortman and Wu, 1996). The most immature T cells their progenitor capacity for NK cells (Ikawa et al., are CD44+CD257 and they develop to the stages of 1999). Therefore, it is conceivable that inactivation of CD44+CD25+, CD447CD25+ and then the E protein activity in bipotent T/NK precursors CD447CD257, after which T cells start to express directs the cell fate to NK cells. In accordance with CD3. By analysing the progenitor activity of fetal this, overexpression of Id3, which has an activity thymocytes at the single cell level using a fetal thymus similar to that of Id2, in immature human fetal organ culture system, Katsura and colleagues demon- hematopoietic cells or thymocytes gives rise to strated that the T/NK bipotent progenitors are found augmented numbers of NK cells at the expense of T mainly in the CD44+CD257 fraction and residual cell development (Heemskerk et al., 1997; Blom et al., bipotent progenitor activity is detected in the 1999). In E2A7/7 mice, however, there is no CD44+CD25+ fraction (Ikawa et al., 1999). Among apparent increase in the NK cell number, (Zhuang, the most immature CD44+CD257 fetal thymocytes, personal communication), although T cell develop- they further identi®ed a CD122 (IL-2Rb)-positive ment is disturbed (Zhuang et al., 1996, 1998; Bain et subfraction that exclusively contains progenitors com- al., 1997). This fact raises the interesting possibility mitted to become NK lineage cells. Analysis of NK that NK cell development requires another transcrip- cells in the fetal thymus of Id27/7 mice has revealed tion factor that positively drives the dierentiation or a role for Id2 in a cell fate decision (Ikawa et al., 2001). that loss of E2A is not equivalent to overexpression of In Id27/7 fetal thymi, thymic NK cells are reduced Id. Defective NK cell development has also been as in the adult organs and Id27/7 fetal thymocytes reported in mice de®cient for Ets-1 (Barton et al., show little dierentiation into NK cells even in 1998) or Stat5b (Imada et al., 1998). These molecules conditions containing cytokines that are highly sup- may be candidates for this positive transcription portive for NK cells in this system. Furthermore, the regulator. In fact, Id proteins have been reported to NK cell progenitor-enriched fraction is missing in interact with Ets family proteins (Yates et al., 1999; Id27/7 thymi. These observations suggest that Ohtani et al., 2001). Id proteins physically interact commitment of fetal thymocytes to the NK cell lineage with the Ets proteins and suppress their functions may be strongly disturbed in Id27/7 mice. In (Yates et al., 1999; Ohtani et al., 2001), whereas NK analyses at the single cell level in fetal thymus organ cell development is impaired in the absence of either cultures, wild-type CD44+CD257CD1227 fetal thy- factor (Barton et al., 1998; Yokota et al., 1999). The mocytes of wild-type embryos are found to contain paradox may be explained by the Ets-Egr1-Id distinct progenitors for T cells and NK cells, in transactivation cascade. Ets proteins are direct up- addition to bipotent T/NK progenitors. In contrast, stream factor of Egr1 (Robinson et al., 1997) As in however, CD44+CD257CD1227 fetal thymocytes of cases of Id1 (Tournay and Benezra, 1996) and Id3 Id27/7 embryos include only progenitors for T cells, (Bain et al., 2001), Id2 may be regulated by Ets-1 via indicating that bipotent T/NK progenitors appear to Egr1 in developing NK cells. acquire the T cell fate in the absence of Id2. Importantly, the total number of progenitors observed Lymphoid organs in this assay is similar in wild-type and Id27/7 embryos implying a qualitative switch in cell fate. The The secondary or peripheral lymphoid organs include expression pro®le of Id2 during early T cell develop- Peyer's patches (PPs) present in the intestine, lymph ment is consistent with these observations. Id2 is nodes (LNs) and the spleen. These are sites for immune expressed in the CD44+CD257 stage and in the reactions brought about by invading microorganisms. CD44+CD257CD122+ subfraction, showing a good The formation of PPs and LNs starts during the fetal correlation with the progenitor activity for NK cells. period and is highly dependent on the lymphotoxin On the other hand, the genes for E proteins such as (LT) signaling pathway through the LTb receptor, E2A and HEB are constantly expressed in all stages of which is triggered by a membrane-bound LTa1/LTb2

Oncogene Id and development Y Yokota 8294 heterotrimer (Fu and Chaplin, 1999). Therefore, any phenotype in Id27/7 or RORg7/7 mice is restricted defect in this signaling pathway can cause a defect in to the generation of the LT-expressing speci®c cell peripheral lymphoid organ development. Examples population during development and the spleens of include LTa7/7 mice (De Togni et al., 1994), these mice are formed normally, although T cells LTb7/7 mice (Koni et al., 1997), LTb receptor- demonstrate increased apoptosis due to impaired de®cient mice (Futterer et al., 1998) and mice bearing a thymic expression of the anti-apoptotic factor Bcl-xL mutation or disruption by targeting of the NF-kB at least in the RORg7/7 mice (Sun et al., 2000). In inducing kinase gene, which is essential for transmitting fact, Id27/7 splenocytes express LT on their surface the signal elicited by the LTb receptor to the nucleus (Honda et al., unpublished observations). (Shinkura et al., 1999; Yin et al., 2001). The LT-expressing cell population involved in the In addition to these mutant mice, there is another formation of PPs and LNs appears to belong to the class of mutants that exhibits a similar alymphoplasia lymphoid lineage and to be derived from immature phenotype, and Id27/7 mice are included in this fetal liver cells with characteristics similar to those of category. This group displays a developmental defect in common lymphoid precursors (Mebius et al., 2001). the generation of cells that play an essential role in the How does Id2 participate in the development of these initiation of peripheral lymphoid organ development unique cells? Defective development of the cells in the by producing membrane-bound LT on their cell absence of Id2 implies that inhibition of the activity of surface. During development at around embryonic bHLH factors is important for the dierentiation of day 14, a unique cell population positive for cell these cells from a progenitor as seen for NK cell surface markers CD4, CD45 and the interleukin-7 development (Ikawa et al., 2001). The responsible receptor a (IL-7Ra), but negative for CD3, is found in factor must play a role in supporting and maintaining peripheral blood and in peripheral lymphoid organs, the main ¯ow of developing cells in the lymphoid but not in the thymus (Mebius et al., 1997). The cells lineage, although it is not clear what kind of bHLH of this population express LT on their cell surface and factor is involved in this process. On the other hand, as have been considered to be crucial for the activation of demonstrated in gene-inactivated mice, RORg is also the genetic program for lymphoid organogenesis required for the development of the LT-expressing (Mebius et al., 1997; Yoshida et al., 1999). In cells. Elucidation of the relationship between Id2 and Id27/7 mice, no PPs or LNs are formed, although RORg will provide us with insights into a molecular the bone marrow, thymus and spleen develop normally event that regulates the development of the unique LT- (Yokota et al., 1999). Flow cytometric analyses fail to expressing cell population during development. We, identify the LT-expressing CD37CD4+IL-7Ra+ cell however, have failed to detect the interaction of these population in the spleen, intestine and peripheral blood molecules at the protein level so far (Karaya et al., of Id27/7 embryos. Consistent with this, colonization unpublished observations) of LT-expressing cells in the developing PPs of wild- type embryos is not observed in the intestine of Id27/7 embryos. These observations strongly indicate The reproductive system that the LT-expressing CD37CD4+IL-7Ra+ cells play an essential role in the organogenesis of peripheral Mammary gland lymphoid organs by expressing LT on the cell surface. Furthermore, this cell population is also missing in The mammary gland is widely studied with respect to mice de®cient for RORg, one of the retinoid receptor- many important aspects of biomedical research includ- related orphan receptors, and the mutant mice lack PPs ing cell dierentiation, proliferation, responses to and LNs (Sun et al., 2000). Interestingly, the splenic hormones, apoptosis and cancer. During pregnancy, microarchitecture is established normally and class the organ exhibits rapid cell proliferation and dier- switching of the immunoglobulins is observed in both entiation into grandular tissues in response to the Id27/7 mice and RORg7/7 mice (Yokota et al., orchestration of hormonal stimuli (Neville and Daniel, 1999; Sun et al., 2000), although reduction of NK cells 1987; Hennighausen and Robinson, 1998). In in RORg mice seems not to be as severe as that Id27/7 female mice, although the mammary gland observed in Id27/7 mice (Littman, personal commu- itself is formed normally, a severe lactation defect is nication). This contrasts with the disorganized T cell/B observed and mammary glands of Id27/7 pregnant cell compartmentalization of the spleen and the mice fail to proliferate and remain immature (Mori et impaired class switching of immunoglobulins which al., 2000). Grafting experiments with the mammary are commonly found in mice defective in LT signaling epithelial tissue clearly indicate that the defect is (Fu and Chaplin, 1999). The spleen is one of the intrinsic to mammary epithelial cells, rather than to peripheral lymphoid organs and requires LT signaling the mesenchymal component or the hormonal function to establish the microarchitecture that is a prerequisite in the mutant mice. Id2 is predominantly expressed in for immune reactions. However, membrane-bound LT mammary epithelial cells and its expression level is provided by CD37CD4+IL77Ra+ cells for the reaches a maximum at around 7 ± 10 days post coitus, formation of PPs and LNs, but by T and B cells for the a time that corresponds well to the stage at which compartmentalization of the spleen (Fu and Chaplin, Id27/7 mammary glands demonstrate the develop- 1999). The defect responsible for the alymphoplasia mental arrest.

Oncogene Id and development Y Yokota 8295 Id2 seems to be important in stimulating cell cycle cycle machinery. This may be related to the myc-Id2- progression of mammary epithelial cells. Id27/7 retinoblastoma protein pathway (Lasorella et al., 2000; mammary epithelial cells show a greatly reduced see also the section by Iavarone in this issue). number of cells in S phase during early pregnancy. This reduction is associated with elevated levels of Spermatogenesis cyclin-dependent kinase inhibitors p27/Kip1 and p21/ WAF1/CIP1. The increased expression of these The testis is a tissue in which each member of the Id suppressors of cell cycle progression may be function- family shows a spatio-temporally distinct expression ally related to the disturbed cell cycle progression in pattern, including speci®c subcellular localization of Id27/7 mammary epithelium during early pregnancy. each protein, which was revealed by an immunohisto- A similar enhancement of p27/Kip1 has been reported chemical study (Sablitzky et al., 1998). There are two in the developing neuroepithelium of compound major cell populations in the testis: one component is Id17/7Id37/7 embryos (Lyden et al., 1999), germ cells, which undergo spermatogenesis, and the implying a common mechanism in the regulation of other is Sertoli cells, somatic cells that constitute the p27/Kip1 expression by Id proteins in these organs. On epithelium of the seminiferous tubules and which the other hand, apoptotic epithelial cells are more nurture germ cells. The process of spermatogenesis is abundant in Id27/7 mice in the late phase of complex and requires multiple steps (Sharpe, 1994). In pregnancy. This is mediated through the p53 ± Bax adult rodents, spermatogonia derived from gonocytes pathway but by a mechanism dierent from that are subdivided into A-type, intermediate and B-type involved in involution of the gland. Although Id2 spermatogonia. A0 spermatogonia are spermatogonial appears to be a survival factor for mammary epithelial stem cells and spermatogenesis starts with their entry cells in the late phase of pregnancy, the molecular into a proliferative phase to produce A1±4 spermato- mechanisms of this have not yet been clari®ed. Both gonia, intermediate spermatogonia and B spermatogo- the proliferation defect and increased apoptosis seem nia through six rounds of cell division. B-type to be responsible for the lactation defect in Id27/7 spermatogonia are cells committed to enter meiosis mice. and they generate the primary (leptotene, zygotene, Another functional role of Id proteins is to suppress pachytene and diplotene) and secondary spermatocytes a bHLH factor that promotes tissue-speci®c cell by the ®rst and second meiotic divisions, respectively. dierentiation (for review see Norton et al., 1998). They further become dividing (MI and MII) sperma- Dysregulation of the factor in the absence of Id2 can tocytes. Postmeiotic spermatid development consists of lead to premature dierentiation (as seen in the central 16 steps in the mouse (Oakberg, 1956). Id1 protein is nervous system of Id17/7Id37/7 mice (Lyden et al., abundantly detected in the cytoplasm of A-type 1999)), which may explain the impaired proliferation of spermatogonia, throughout MI and MII spermatocytes Id27/7 mammary epithelial cells. However, this is and in the cytoplasm of spermatids at the mid-late less likely to be a cause of the lactation defect. stages of development. Id4 is expressed in the Id27/7 mammary epithelial cells remain in an cytoplasm of A-type spermatogonia and shows nuclear immature state from the standpoint of expression of or perinuclear expression in the late pachytene milk protein. This could on the other hand simply spermatocytes and spermatids. On the other hand, re¯ect the fact that these cells are being propelled down Id2 expression is primarily detected in the nucleus of a `wrong' pathway and do not express the expected Sertoli cells, together with a spatio-temporally complex markers. The putative tissue-speci®c bHLH factor expression in germ cells in the meiotic prophase and involved in mammary gland development has not been postmeiotic phase. Id3 is detected restrictedly in the identi®ed yet (Boku and Yokota, unpublished observa- cytoplasm of Sertoli cells. These facts raise the tions). interesting possibility that the four members of the Many genes other than Id2 are also required to family have distinct roles in spermatogenesis. achieve the normal development of mammary glands In adult Id27/7 mice, the seminiferous tubules are during pregnancy (Hennighausen and Robinson, 1998). abnormal and show a dramatic decrease in mature These include prolactin (Horseman et al., 1997), sperm. The number of Sertoli cell number is reduced prolactin receptor (Ormandy et al., 1997), Stat5a (Liu and apoptosis is increased. It seems, therefore, that Id2 et al., 1997), progesterone receptor (Lydon et al., function in Sertoli cells is crucial for normal 1995), cyclin D1 (Sicinski et al., 1995; Fantl et al., spermatogenesis (Sablitzky, personal communication). 1995), C/EBPb (Seagroves et al., 1998) and A-myb This phenotype is consistent with the expression (Toscani et al., 1997). None of these genes show pattern of Id2 and suggests a speci®c role for Id2 in reduced expression in the mammary glands of pregnant Sertoli cells (Sablitzky et al., 1998). What kind of Id27/7 mice, suggesting that Id2 functions as a function does Id2 exert in Sertoli cells? Sertoli cells downstream or parallel factor to these molecules in the proliferate during fetal and neonatal periods but the cascade regulating the development of lactating rate of proliferation drops substantially before puberty mammary glands. Considering the defective cell cycle (Russell et al., 1990). This implies that the number of progression in Id27/7 mammary epithelial cells, it is Sertoli cells is constant afterwards, so long as they highly probable that Id2 acts at a point where survive during adulthood. It is therefore highly extracellular signals converge and turns on the cell probable that Id2 is needed for the long-term survival

Oncogene Id and development Y Yokota 8296 of Sertoli cells in the adult testes. Although the (a) molecular mechanisms of this function are still unclear, further analyses may provide new insight into an unknown function of Id2 in non- or slowly dividing cells. Furthermore, such analyses may explain why Id2 is highly expressed in postmitotic neurons such as cortical neurons in the cerebrum and Purkinje cells in the cerebellum. It is tempting to speculate that the function of Id2 in Sertoli cells is related to that in mammary epithelial cells in the late phase of pregnancy, where increased apoptosis is encountered in the absence of Id2 (Mori et al., 2000).

Concluding remarks

As discussed above, Id proteins are involved in the development of diverse organs and tissues. Two biological characteristics of Id proteins, inhibition of dierentiation and stimulation of proliferation, were (b) mainly discovered by using in vitro model systems (for review see Norton et al., 1998). These activities have been veri®ed in vivo as well. Although it is dicult to separate these two functions clearly, one or the other of these two functions of the Id proteins may be more prominent depending on the particular cell type. In Id27/7 mice, for example, defects in lactation and NK cell development are ascribable primarily to (c) proliferation disorder of mammary epithelial cells and failure in cell fate determination to NK cells, respectively. The activity of Id proteins as regulators of bHLH factors can be further divided into two categories: one- Figure 1 Functions of Id proteins in vivo. The dierent modes of way and bifurcation types (Figure 1a). An example of the functions of Id proteins; control of dierentiation, prolifera- the one-way type is neuronal dierentiation, as seen in tion and survival, are schematically represented in (a), (b) and (c), the Id17/7Id37/7 mouse brain (Lyden et al., 1999). respectively. (a) The one-way type of dierentiation control is the Id proteins are required as inhibitors of dierentiation classical model. A bHLH factor promotes cell dierentiation and Id inhibits it by suppressing the function of the bHLH factor. The to maintain the immaturity of neuroblasts and permit bifurcation type is seen in the case of NK cell development. A their proliferation until some appropriate time point bHLH factor stimulates dierentiation of cell type X and during development. In the absence of Id proteins, suppresses the other pathway to cell type Y. Id promotes bHLH factors are expressed or start to work dierentiation of cell type Y by suppressing the activity of the bHLH factor, which inhibits dierentiation to cell type X. Note prematurely and the terminal dierentiation is im- that Id is an inhibitor for cell type X but a stimulator for cell type properly induced earlier than normal. The bifurcation Y. (b) Id maintains the immaturity of cells and stimulates their type is seen in NK cell development (Ikawa et al., proliferation. (c) Id functions as a survival factor for postmitotic 2001). Bipotent T/NK precursors give rise to T cells as dierentiated cells. In (b) and (c), it is not clear whether bHLH well as NK cells. T cell development is driven by factors are involved bHLH factors, and the inactivation of these factors leads to cell fate speci®cation to the other pathway to NK cell development. Thus, Id proteins are indeed inhibitors of dierentiation for T cells but, inversely, (Roberts et al., 2001), Elk-1 (Yates et al., 1999), c-Ets-1 stimulators of dierentiation for NK cells. On the (Ohtani et al., 2001) and the retinoblastoma protein other hand, Id proteins are essential for the prolifera- family (Iavarone et al., 1994; Lasorella et al., 1996, tion of pregnant mammary epithelial cells (Mori et al., 2000). Elucidation of the detailed relationships of Id 2000) and the long-term survival of Sertoli cells after proteins with these factors would reveal novel roles for the cessation of cell division (Sablitzky, personal Id proteins in dierentiation and proliferation controls, communications) (Figure 1b,c). and promote our understanding of how dierentiation In addition to the progress in the Id-bHLH system, and proliferation are regulated. Eleven years after the recent studies have revealed that Id proteins can discovery of Id (Benezra et al., 1990), we have just regulate a variety of proteins other than bHLH factors begun to gather the abundant fruits of Id research. at the protein level. These include MIDA1 (Shoji et al., Since the work is still expanding in many directions, 1995; Inoue et al., 1999, 2000), Pax2, Pax5 and Pax8 one can be optimistic about the future.

Oncogene Id and development Y Yokota 8297 Acknowledgments Biochemistry, Fukui Medical University, for helpful dis- IamgratefultoFredSablitzky,DanLittmanandYuan cussions and comments on the manuscript. I also thank Y Zhuang for sharing their unpublished observations and to Matsui for her secretarial assistance. all members of my laboratory at the Department of

References

Andres-Barquin PJ, Hernandez MC and Israel MA. (2000). Inoue T, Shoji W and Obinata M. (1999). Biochem. Biophys. Histol. Histopathol., 15, 603 ± 618. Res. Commun., 266, 147 ± 151. Bain G, Cravatt CB, Loomans C, Alberola-Ila J, Hedrick Inoue T, Shoji W and Obinata M. (2000). Genes Cells, 5, SM, Murre C. (1997). Nat. Immunol., 2, 165 ± 171. 699 ± 709. Bain G, Engel I, Robanus Maandag EC, te Riele HPJ, Jen Y, Weintraub H and Benezra R. (1992). Genes Dev., 6, Voland JR, Sharp LL, Chun J, Huey B, Pinkel D and 1466 ± 1479. Murre C. (1997). Mol. Cell. Biol., 17, 4782 ± 4791. Jen Y, Manova K and Benezra R. (1996). Dev. Dyn., 207, Barres BA, Lazar MA and Ra MC. (1994). Development, 235 ± 252. 120, 1097 ± 1108. Jen Y, Manova K and Benezra R. (1997). Dev. Dyn., 208, Barton K, Muthusamy N, Fischer C, Ting CN, Walunas TL, 92 ± 106. Lanier L and Leiden JM. (1998). Immunity, 9, 555 ± 563. Kim D, Peng XC and Sun XH. (1999). Mol. Cell. Biol., 19, Benezra R, Davis RL, Lockshon D, Turner DL and 8240 ± 8253. Weintraub H. (1990). Cell, 61, 49 ± 59. Kondo T and Ra M. (2000a). EMBO J., 19, 1998 ± 2007. Blom B, Heemskerk MHM, Verschuren MCM, van Dongen Kondo T and Ra M. (2000b). Development, 127, 2989 ± JJM, Stegman APA, Bakker AQ, Couwenberg F, Res 2998. PCM and Spits H. (1999). EMBO J., 18, 2793 ± 2802. Koni PA, Sacca R, Lawton P, Browning JL, Ruddle NH and Cai L, Morrow EM and Cepko CL. (2000). Development, Flavell RA. (1997). Immunity, 6, 491 ± 500. 127, 3021 ± 3030. Lasorella A, Iavarone A and Israel MA. (1996). Mol. Cell. Carlyle JR and Zuniga-P¯ucker JC. (1998). Immunol. Rev., Biol., 16, 2570 ± 2578. 165, 63 ± 74. Lasorella A, Noseda M, Beyna M, Yokota Y and Iavarone CarsonWE,GiriJG,LindemannMJ,LinettML,AhdiehM, A. (2000). Nature, 407, 592 ± 598. Paxton R, Anderson D, Eisenmann J, Grabstein K and Lee JE. (1997). Curr. Opin. Neurobiol., 7, 13 ± 20. Caligiuri MA. (1994). J. Exp. Med., 180, 1395 ± 1403. Liu X RG, Wagner KU, Garrett L, Wynshaw-Boris A and Christy BA, Sanders LK, Lau LF, Copeland NG, Jenkins Hennighausen L. (1997). Genes Dev., 11, 179 ± 186. NA and Nathans D. (1991). Proc.Natl.Acad.Sci.USA, LuQR,YukD,AlbertaJA,ZhuZ,PawlitzkyI,ChanJ, 88, 1815 ± 1819. McMahon AP, Stiles CD and Rowitch DH. (2000). De Togni P, Goellner J, Ruddle NH, Streeter PR, Fick A, Neuron, 317 ± 329. Mariathasan S, Smith SC, Carlson R, Shornick LP, LydenD,YoungAZ,ZagzagD,YanW,GeraldW,O'Reilly Strauss Schoenberger J, Russell JH, Karr R, Chaplin R, Bader BL, Hynes RO, Zhuang Y, Manova K and DD. (1994). Science, 264, 703 ± 707. Benezra R. (1999). Nature, 401, 670 ± 677. Duncan M, DiCicco Bloom EM, Xiang X, Benezra R and Lydon JP, DeMayo FJ, Funk CR, Mani SK, Hughes AR, Chada K. (1992). Dev. Biol., 154, 1±10. Montgomery CAJ, Shyamala G, Conneely OM and Duncan MK, Bordas L, Dicicco Bloom E and Chada KK. O'Malley BW. (1995). Genes Dev., 15, 2266 ± 2278. (1997). Dev. Dyn., 208, 107 ± 114. Martinsen BJ and Bronner Fraser M. (1998). Science, 281, Durand B and Ra M. (2000). Bioessays, 22, 64 ± 71. 988 ± 991. Evans SM and O'Brien TX. (1993). Dev. Biol., 159, 485 ± 499. Massari ME and Murre C. (2000). Mol. Cell. Biol., 20, 429 ± FantlV,StampG,AndrewsA,RosewellIandDicksonC. 440. (1995). Genes Dev., 9, 2364 ± 2372. Mebius RE, Rennert P and Weissman IL. (1997). Immunity, Fu Y and Chaplin DD. (1999). Annu. Rev. Immunol., 17, 7, 493 ± 504. 399 ± 433. Mebius RE, Miyamoto T, Christensen J, Domen J, Cupedo Futterer A, Mink K, Luz A, Kosco-Vilbois MH and Pfeer T, Weissman IL and Akashi K. (2001). J. Immunol., 166, K. (1998). Immunity, 9, 59 ± 70. 6593 ± 6601. Heemskerk MH, Blom B, Nolan G, Stegmann AP, Bakkar Mori S, Nishikawa S-I and Yokota Y. (2000). EMBO J., 19, AQ, Weijer K, Res PC, Spits H. (1997). J. Exp. Med., 186, 5772 ± 5781. 1597 ± 1602. Morrow MA, Mayer EW, Perez CA, Adlam M and Siu G. Hennighausen L and Robinson GW. (1998). Genes Dev., 12, (1999). Mol. Immunol., 36, 491 ± 503. 449 ± 455. Neuman T, Keen A, Zuber MX, Kristjansson GI, Gruss P Horseman ND, Zhao W, Montecino-Rodriguez E, Tanaka and Nornes HO. (1993). Dev. Biol., 160, 186 ± 195. M, Nakashima K, Engle SJ, Smith F, Marko E and Neville MC and Daniel CW. (1987). The mammary gland: Dorshkind K. (1997). EMBO J., 16, 6926 ± 6935. development, regulation, and function. Plenum Press, New Iavarone A, Garg P, Lasorella A, Hsu J and Israel MA. York. (1994). Genes Dev., 8, 1270 ± 1284. Norton JD, Deed RW, Craggs G and Sablitzky F. (1998). Ikawa T, Kawamoto H, Fujimoto S and Katsura Y. (1999). Trends Cell Biol., 8, 58 ± 65. J. Exp. Med., 190, 1617 ± 1626. Oakberg EF. (1956). Am.J.Anat.,99, 507 ± 516. Ikawa T, Fujimoto S, Kawamoto H, Katsura Y and Yokota Ohtani N, Zebedee Z, Huot TJ, Stinson JA, Sugimoto M, Y. (2001). Proc. Natl. Acad. Sci. USA, 98, 5164 ± 5169. Ohashi Y, Sharrocks AD, Peters G and Hara E. (2001). Imada K, Bloom ET, Nakajima H, Horvath-Arcidiacono Nature, 409, 1067 ± 1070. JA, Udy GB, Davey HW and Leonard WJ. (1998). J. Exp. Olson EN and Klein WH. (1994). Genes Dev., 8, 1±8. Med., 188, 2067 ± 2074.

Oncogene Id and development Y Yokota 8298 Ormandy CJ, Camus A, Barra J, Damotte D, Lucas B, Sun XH. (1994). Cell, 79, 893 ± 900. Buteau H, Edery M, Brousse N, Babinet C, Binart N and Sun Z, Unutmaz D, Zou YR, Sunshine MJ, Pierani A, Kelly P.A. (1997). Genes Dev., 15, 167 ± 178. Brenner-Morton S, Mebius RE and Littman DR. (2000). Riechmann V, van Cruchten I and Sablitzky F. (1994). Science, 288, 2369 ± 2373. Nucleic Acids Res., 22, 749 ± 755. Temple S and Ra MC. (1986). Cell, 14, 773 ± 779. Riechmann V and Sablitzky F. (1995). Cell Growth Dier., 6, Toscani A, Mettus RV, Coupland R, Simpkins H, Litvin J, 837 ± 843. Orth J, Hatton KS and Reddy EP. (1997). Nature, 386, Rivera RR, Johns CP, Quan J, Johnson RS and Murre C. 713 ± 717. (2000). Immunity, 12, 17 ± 26. Tournay O, Benezra R. (1966). Mol. Cell. Biol., 16, 2418 ± Roberts EC, Deed RW, Inoue T, Norton JD and Sharrocks 2430. AD. (2001). Mol. Cell. Biol., 21, 524 ± 533. Tzeng SF and de Vellis J. (1998). Glia, 24, 372 ± 381. Robinson L, Panayiotakis A, Papas TS, Kola I, Seth A. Wang S, Sdrulla A, Johnson JE, Yokota Y and Barres BA. (1997). Proc. Natl. Acad. Sci. USA, 94, 7170 ± 7175. (2001). Neuron, 29, 603 ± 614. Russell LD, Ettlin RA, Hickim APS and Clegg ED. (1990). Wice BM and Gordon JI. (1998). J. Biol. Chem., 273, 25310 ± Histological and histopathological evaluation of the testis. 25319. Cache River Press. Williams NS, Klem J, Puzanov IJ, Sivakumar PV, Schatzle Sablitzky F, Moore A, Bromley M, Deed RW, Newton JS JD, Bennett M and Kumar V. (1998). Immunol. Rev., 165, and Norton JD. (1998). Cell Growth Dier., 9, 1015 ± 1024. 47 ± 61. Seagroves TN, Krnacik S, Raught B, Gay J, Burgess-Beusse Yan W, Young AZ, Soares VC, Kelley R, Benezra R and B, Darlington GJ and Rosen JM. (1998). Genes Dev., 15, Zhuang Y. (1997). Mol. Cell. Biol., 17, 7317 ± 7327. 1917 ± 1928. Yates PR, Atherton GT, Deed RW, Norton JD and Sharpe RM. (1994). Physiology of Reproduction, Vol. 1. Sharrocks AD. (1999). EMBO J., 18, 968 ± 976. Knobil E and Neil, JD (eds). Raven Press: New York, YinL,WuL,WeschH,ArthurCD,WhiteJM,GoeddelDV pp. 1363 ± 1434. and Schreiber RD. (2001). Science, 291, 2162 ± 2165. Shinkura R, Kitada K, Matsuda F, Tashiro K, Ikuta K, Yokota Y, Mansouri A, Mori S, Sugawara S, Adachi S, Suzuki M, Kogishi K, Serikawa T and Honjo T. (1999). Nishikawa S-I and Gruss P. (1999). Nature, 397, 702 ± 706. Nat. Genet., 22, 74 ± 77. Yoshida H, Honda K, Shinkura R, Adachi S, Nishikawa S, Shoji W, Inoue T, Yamamoto T and Obinata M. (1995). J. Maki K, Ikuta K and Nishikawa S-I. (1999). Int. Biol. Chem., 270, 24818 ± 24825. Immunol., 11, 643 ± 655. Shortman K and Wu L. (1996). Ann. Rev. Immunol., 14, 29 ± Zhou Q, Wang S and Anderson DJ. (2000). Neuron, 25, 331 ± 47. 343. SicinskiP,DonaherJL,ParkerSB,LiT,FazeliA,Gardner Zhu W, Dahmen J, Bulfone A, Rigolet M, Hernandez MC, H, Haslam S, Bronson RT, Elledge SJ and Weinberg RA. Kuo WL, Puelles L, Rubenstein JL and Israel MA. (1995). (1995). Cell, 82, 621 ± 630. Brain Res. Mol. Brain Res., 30, 312 ± 326. SpitsH,BlomB,JalecoAC,WeijerK,VerschurenMCM, Zhuang Y, Cheng P and Weintraub H. (1996). Mol. Cell. van Dongen JJM, Heemskerk MHM and Res PCM. Biol., 16, 2898 ± 2905. (1998). Immunol. Rev., 165, 75 ± 86. Zhuang Y, Barndt RJ, Pan L, Kelley R and Dai M. (1998). Sun XH, Copeland NG, Jenkins NA and Baltimore D. Mol. Cell. Biol., 18, 3340 ± 3349. (1991). Mol. Cell. Biol., 11, 5603 ± 5611.

Oncogene