Expression of Islet Homologs in Chickens and Mammals: Brain and Spinal Cord Motoneurons

Motor neurons located at different positions in the embryonic spinal cord innervate distinct targets in the periphery, establishing a topographic neural map. The topographic organization of motor projections depends on the generation of subclasses of motor neurons that select specific paths to their targets. A family of LIM homeobox genes has been cloned in the chick. The combinatorial expression of four of these genes, Islet-1, Islet-2, Lim-1, and Lim-3, defines subclasses of motor neurons that segregate into columns in the spinal cord and select distinct axonal pathways. Thus the combination of LIM domain proteins serve to code motor neuron identity in the spinal cord (Tsuchida, 1994).

Sonic hedgehog (Shh) is strongly implicated in the development of ventral structures in the nervous system. Addition of Sonic hedgehog protein to chick spinal cord explants induces floor plate and motoneuron development. Whether Shh acts directly to induce these cell types or whether their induction is mediated by additional factors is unknown. To further investigate the role of Shh in spinal neuron development, low-density cultures of murine spinal cord precursor cells were used. Shh stimulates neuronal differentiation; however, it does not increase the proportion of neurons expressing the first postmitotic motoneuron marker Islet-1. Moreover, Shh induces Islet-1 expression in neural tube explants, suggesting that it acts in combination with neural tube factors to induce motoneurons. Another factor implicated in motoneuron development is neurotrophin 3 (NT3): when assayed in isolated precursor cultures, it has no effect on Islet-1 expression. However, the combination of N-terminal Shh and NT3 induces Islet-1 expression in the majority of neurons in low-density cultures of caudal intermediate neural plate. Furthermore, in explant cultures, Shh-mediated Islet-1 expression is blocked by an anti-NT3 antibody. Previous studies have shown expression of NT3 in the region of motoneuron differentiation and that spinal fusimotor neurons are lost in NT3 knock-out animals. Taken together, these findings suggest that Shh can act directly on spinal cord precursors to promote neuronal differentiation, but induction of Islet-1 expression is regulated by factors additional to Shh, including NT3 (Dutton, 1999).

Different neuronal subpopulations derived from in vitro differentiation of embryonic stem (ES) cells have been characterized using as markers the expression of several homeodomain transcription factors. Following treatment of embryo-like aggregates with retinoic acid (RA), Pax-6, a protein expressed by ventral central nervous system (CNS) progenitors, is induced. In contrast, Pax-7 expressed in vivo by dorsal CNS progenitors, and erbB3, a gene expressed by neural crest cells and its derivatives, are almost undetectable. CNS neuronal subpopulations generate expressed combination of markers characteristic of somatic motoneurons (Islet-1/2, Lim-3, and HB-9); cranial motoneurons (Islet-1/2 and Phox2b) and interneurons (Lim-1/2 or EN1). Molecular characterization of neuron subtypes generated from ES cells should considerably facilitate identification of new genes expressed by restricted neuronal cell lineages (Renoncourt, 1998).

These LIM homeodomain proteins are expressed prior to the formation of distinct motor axon pathways and before motor columns appear. Depending on their arrangement in columns and eventual synaptic targets, motor neurons of the chick brain stem are designated as belonging to somatic motor (sm) visceral motor (vm), or branchiomotor (bm) classes or to the ipsilateral or contralateral vestibuloacoustic effect neuronal population. Sm neurons innervate muscle derived from the paraxial mesoderm and prechordal plate mesoderm. Bm, vm and vestibuloacoustic axons extend dorsolaterally for some distance through the neuroepithelium before converging on large single exit points within the dorsal neural tube (alar plate). Bm neurons innervate muscle derived from paraxial mesoderm within the branchial arches, while vm neurons innervate parasympathetic ganglia associated with lacrimal and salivary glands or neuronal plexuses that innervate smooth muscle; vestibuloacoustic efferent neurons innervate the hair cells of the inner ear. Subpopulations of spinal motor neurons within specific locations in the spinal cord and distinct targets in the periphery express different combinations of LIM homeobox genes. Sm neurons of the medial division of the median motor column express Islet-1, Isl-2, and Lim-3, while those of the lateral division of the median motor column and the medial division of the later motor column express Isl-1 and Isl-2. Sm neurons of the lateral division of the lateral motor column express Lim-1 and Isl-2. Since the lateral motor column is present only at limb levels, Lim-1 expression is restricted to these levels of the neuraxis. At early stages, visceral motor neurons express both Isl-1 and Isl-2, but after their migration to form the column of Terni, only a subset of these neurons continues to express Isl-1. These genes are good candidates to confer target specificity upon motor neuron classes, since they are expressed at times before the motor columns have fully segregated and before axons have reached their targets (Tsuchida, 1994).

Motor neuron differentiation in the mouse is accompanied by the expression of a LIM homeodomain transcription factor, Islet1 (ISL1). Motor neurons are not generated without ISL1, although many other aspects of cell differentiation in the neural tube occur normally. A population of interneurons that express Engrailed1 (Drosophila homolog: Engrailed), however, also fails to differentiate in Isl1 mutant embryos. The differentiation of EN1+ interneurons can be induced in both wild-type and mutant neural tissue by regions of the neural tube that contain motor neurons. These results show that ISL1 is required for the generation of motor neurons and suggest that motor neuron generation is required for the subsequent differentiation of certain EN1 expressing interneurons (Pfaff, 1996).

The generation of distinct classes of motor neurons is an early step in the control of vertebrate motor behavior. To study the interactions that control the generation of motor neuron subclasses in the developing avian spinal cord in vivo grafting studies were performed in which either the neural tube or flanking mesoderm were displaced between thoracic and brachial levels. The positional identity of neural tube cells and motor neuron (MN) subtype identity was assessed by Hox and LIM homeodomain protein expression. Brachial (B) levels of the median motor column (MMC) are organized into three columns: neurons of the medial MMC (MMCM) co-express Isl1, Isl2 and Lim3, neurons of the medial lateral motor column (LMCM) co-express Isl1 and Isl2, and motoneurons of the lateral LMC (LMCL) coexpress Isl2 and Lim1. At thoracic (T) levels motoneurons are also organized into three columns: MMCM neurons; lateral MMC neurons that coexpress Isl1 and Isl2 but not Lim3, and dorsomedially positioned Column of Terni (CT) neurons that express only Isl1. Grafts of 13-15 segment quail T neural tube were placed rostrally at the B level of 12-15 segment chick hosts. Marker and morphological analysis reveals that grafted neural cells divert their normal T fates and their neuronal progeny acquire the molecular properties of B MNs. These changes in the neural tube are restricted to a limited time frame. The rostrocaudal identity of neural cells is plastic at the time of neural tube closure and is sensitive to positionally restricted signals from the paraxial mesoderm. Such paraxial mesodermal signals appear to control the rostrocaudal identity of neural tube cells and the columnar subtype identity of motor neurons. Analysis of neural Hoxc8 expression provides evidence that the change in cell identity after neural tube displacement is not restricted to the MNs; the change occurs in a graded manner along the rostrocaudal axos of the spinal cord, and is associated with both a rostral and caudal respecification in cell fate. In contrast, neural tube grafts between B and T levels do not change the pattern of Hoxc8 expression in the flanking paraxial mesodem. These results suggest that the generation of motor neuron subtypes in the developing spinal cord involves the integration of distinct rostrocaudal and dorsoventral patterning signals that derive, respectively, from paraxial and axial mesodermal cell groups (Ensini, 1998).

The diversification of neuronal cell types in the vertebrate central nervous system depends on inductive signals provided by local organizing cell groups of both neural and nonneural origin. The link between neuronal birth date, migratory pattern, and identity is also evident in the generation of motor neurons in the spinal cord. These conserved features are particularly apparent for motor neurons of the lateral motor column (LMC). This class of motor neurons is generated selectively at brachial and lumbar levels of the spinal cord, and their axons innervate target muscles in the limb. Within the LMC, motor neurons can be further divided into two subclasses: medial LMC neurons that project to ventrally derived limb muscles, and lateral LMC neurons that project to dorsally derived limb muscles. Motor neurons destined to form the medial LMC leave the cell cycle before lateral LMC neurons; as a consequence, prospective lateral LMC neurons emerge from the ventricular zone and migrate past medial LMC neurons to their final position. The time of generation and the distinct migratory environment represent two prominent differences between the development of lateral LMC neurons and other motor neurons. In addition, the total number of motor neurons generated at limb levels of the spinal cord is greater than that at nonlimb levels, presumably to accommodate the formation of the LMC (Sockanathan, 1998 and references).

All somatic motor neurons initially express Isl1 and Isl2, and most maintain the expression of these genes. Lateral LMC neurons, however, extinguish Isl1 and initiate Lim1 expression as they begin to migrate past medial LMC neurons, thus acquiring a unique LIM homeobox gene code. Studies of LIM homeobox gene function in vertebrates and invertebrates have provided evidence that this gene family has a role in motor neuron differentiation and axon pathfinding. The diversification of motor neuron subtypes is initiated by inductive signals from the axial and paraxial mesoderm that operate along the dorsoventral and rostrocaudal axes of the neural tube. However, medial and lateral LMC motor neurons are generated from progenitor cells that occupy the same dorsoventral and rostrocaudal positions, and thus it is unlikely that mesodermal signals impose this distinction. The late birth date of lateral LMC neurons and their migration past early-born LMC neurons prompted a consideration of whether the fate of lateral LMC neurons might be directed by signals provided by early-born LMC neurons. This hypothesis invokes the idea that LMC motor neurons generated at early stages express a local but non-cell-autonomous signal that induces the lateral LMC phenotype in late-born LMC neurons. A retinoid-mediated signal provided by one subset of early-born spinal motor neurons (the medial) imposes a local variation in the number of motor neurons generated at different axial levels and also specifies the identity of a later-born subset of motor neurons (the lateral). Thus, in the vertebrate central nervous system the distinct fates of late-born neurons may be acquired in response to signals provided by early-born neurons (Sockanathan, 1998).

To begin to define the contribution of retinoid signaling to motor neuron differentiation, the pattern of expression of retinaldehyde dehydrogenase 2 (RALDH2) in the developing spinal cord was examined. At brachial levels, RALDH2 expression is first detected at stage 19, and at this and subsequent stages, expression in the ventral spinal cord appears to be restricted to motor neurons. By stage 27, when the medial motor column (MMC) and LMC have segregated, expression of RALDH2 is restricted to LMC neurons. Within the LMC, RALDH2 is expressed by both medial and lateral LMC neurons. A similar LMC-specific pattern of RALDH2 expression is detected at lumbar levels. Consistent with the restriction of RALDH2 expression to LMC neurons, no expression of the gene is detected in motor neurons at thoracic levels. The expression of RALDH2 in motor neurons at brachial and lumbar levels persists until at least stage 35, although from stage 29 onward, expression gradually becomes restricted to specific motor neuron pools. The only other site of RALDH2 expression in the spinal cord is in the roof plate, both at limb and nonlimb levels. These selective results show that (1) RALDH2 expression is initiated during the early phase of motor neuron generation at brachial levels of the spinal cord; (2) RALDH2 expression distinguishes developing LMC neurons from other somatic or visceral motor neurons, and (3) RALDH2 expression precedes the appearance of Isl2+, Lim1+ lateral LMC neurons (Sockanathan, 1998).

The number of Isl+ motor neurons was counted in brachial ventral/floor plate (vf) explants grown either alone or with retinol (Rol), a metabolic precursor of retinoic acid, or with all-trans retinoic acid (RA). The number of Isl+ motor neurons in [vf] explants grown in the presence of either Rol or RA is increased by 60%. The detection of an increase in motor neuron number with Rol, as well as with RA, indicates that explants grown in medium with no added retinoid are deprived of the metabolic substrate required for synthesis of RA by RALDH2. To examine further the involvement of RALDH2 activity in the control of motor neuron number, thoracic [vf] explants, which do not express RALDH2, were exposed to Rol or RA and the number of Isl+ motor neurons measured. In contrast to results obtained with brachial level explants, exposure of thoracic [vf] explants to Rol does not increase motor neuron number, whereas RA similarly induces a 60% increase in Isl+ motor neurons. Taken together, these results provide evidence that (1) retinoids increase the number of motor neurons; (2) the increase in motor neuron number detected after exposure of brachial [vf] explants to Rol is correlated with the synthesis of active retinoids by RALDH2 activity, and (3) the apparent requirement for RALDH2-generated retinoids can be overcome by exogenous RA. The retinoid-induced increase in motor neuron number at brachial levels appears to result from an increase in the number of progenitor cells. These experiments suggest that, at limb levels, a RALDH2-generated LMC source of retinoids acts non-cell-autonomously to increase the number of motor neuron progenitors and, consequently, postmitotic motor neurons. Studies using an RAR antagonist show that retinoid receptor activation is required for the generation of lateral LMC neurons and for the control of motor neuron number. Maintenance of the lateral LMC phenotype appears to require ongoing retinoid signaling over the period that these neurons are migrating to their lateral position (Sockanathan, 1998).

RALDH2-dependent induction of lateral LMC neurons requires non-autonomous RA signaling. The onset of RALDH2 and Lim1 expression by lateral LMC neurons was examined. At stage 23, many Isl2+, Lim1+ lateral LMC neurons are still located medial to Isl1+, Isl2+ medial LMC neurons. These Isl2+, Lim1+ neurons do not express RALDH2, suggesting that their lateral LMC phenotype has not been acquired through cell-intrinsic RALDH2 activity. Many of the motor neurons that are located in an even more medial position, distant from RALDH2+ neurons, will populate the lateral LMC, but at this stage these neurons express Isl1/2 but not Lim1. These observations support the idea that the lateral phenotype of LMC neurons is acquired by virtue of the proximity of the neurons to a RALDH2-dependent signal provided by earlier-born LMC neurons. The late birth date of lateral LMC neurons requires that they migrate past early-born neurons to reach their final position. What role might this inside-out program of neuronal migration have in the establishment of the lateral LMC phenotype? The detection of late-born Isl2+, Lim1+ lateral LMC neurons in positions adjacent but medial to early-born RALDH2+ medial LMC neurons provides evidence that proximity to early-born neurons is sufficient to achieve a lateral LMC identity. The failure of late-born LMC neurons to migrate past medial LMC neurons might, however, have the consequence that some LMC neurons fail to be exposed to retinoid signals before they lose competence to respond. In this view, the migration of prospective lateral LMC neurons through early-born LMC neurons would achieve a rapid intermixing of inductive and responsive neurons and ensure that the entire population of late born LMC neurons efficiently encounters a local source of retinoid signals (Sockanathan, 1998).

In mammals, Pax-6 (Drosophila homolog: Eyeless)is expressed in several discrete domains of the developing CNS and has been implicated in neural development, although its precise role remains elusive. A novel Small eye rat strain (rSey2) was found with phenotypes similar to mouse and rat Small eye. Analyses of the Pax-6 gene reveals one base (C) insertion in an exon encoding the region downstream of the paired box of the Pax-6 gene (resulting in the rSey2 mutation), resulting in the generation of a truncated protein due to the frame shift. rSey2/rSey2 mutant rats exhibit abnormal development of motor neurons in the hindbrain. The Islet-1-positive motor neurons are generated just ventral to the Pax-6-expressing domain, both in the wild-type and mutant embryos. However, two somatic motor (SM) nerves, the abducens and hypoglossal nerves, are missing in homozygous embryos. No SM-type axonogenesis (ventrally growing) is found in the mutant postotic hindbrain, though branchiomotor and visceral motor (BM/VM)-type axons (dorsally growing) are observed within the neural tube. To discover whether the identity of these motor neuron subtypes is changed in the mutant, expression of LIM homeobox genes Islet-1, Islet-2 and Lim-3 were examined. At the postotic levels of the hindbrain, SM neurons express all the three LIM genes, whereas BM/VM-type neurons are marked by Islet-1 only. In the Pax-6 mutant hindbrain, Islet-2 expression is specifically missing, which results in the loss of the cells harboring the post-otic hindbrain SM-type LIM code (Islet-1 + Islet-2 + Lim-3). Expression of Wnt-7b, which overlaps with Pax-6 in the ventrolateral domain of the neural tube, is also specifically missing in the mutant hindbrain, while it remained intact in the dorsal non-overlapping domain. These results strongly suggest that Pax-6 is involved in the specification of subtypes of hindbrain motor neurons, presumably through the regulation of Islet-2 and Wnt-7b expression. Since Islet-2 and Pax-6 expression domains do not overlap, is suggested that Islet-2 expression is regulated by indirecly by Pax-6 via Wnt-7b. This makes evolutionary sense as it is known that Drosophila paired regulates wingless and Wnt-1 expression in the midbrain/hindbrain boundary is controlled by Pax-2,5, 8. Wnt-7b may be involved in either specification of neuronal subdypes or axon pathfinding (Osumi, 1997).

Hox genes have been implicated in specifying positional values along the anteroposterior axis of the caudal central nervous system, but their nested and overlapping expression has complicated the understanding of how they confer specific neural identity. A direct gain-of-function approach was employed using retroviral vectors to misexpress Hoxa2 and Hoxb1 outside of the normal Hox expression domains, thereby avoiding complications resulting from possible interactions with endogenous Hox genes. Misexpression of either Hoxa2 or Hoxb1 in the anteriormost hindbrain (rhombomere1, r1) leads to the generation of motor neurons in this territory, even though this region is normally devoid of this cell type. Depending on the target tissue they innervate, motor neurons in the hindbrain can be classified into three different subtypes: visceromotor, branchiomotor and somatomotor. An attempt was made to determine whether misexpression of Hoxb1 leads to the induction of a particular subtype of motor neuron or rather to the induction of a generic type of motor neuron. The migratory behaviour displayed by the ectopic motor neurons in r1 indicate that they might be of the branchiomotor subtype. To test this at the molecular level, use was made of the observation that in the hindbrain only somatomotor neurons but not branchiomotor neurons express the LIM-homeobox gene Isl2. When analysed at HH25 stage, somatic trochlear and abducens motor neurons coexpress Isl1 and Isl2, whereas branchiomotor neurons express Isl1 but not Isl2. Ectopic motor neurons in r1 are never found to express Isl2 in addition to Isl1, consistent with their being of the branchiomotor subtype. This result demonstrated, therefore, that the activity of Hoxb1 was sufficient to selectively induce the generation of a distinct specified subtype of motor neurons (Jungbluth, 1999).

In the case of Hoxb1-induced cells, their axons leave the hindbrain either by fasciculating with the resident cranial motor axons at isthmic (trochlear) or r2 (trigeminal) levels of the axis or via novel ectopic exit points in r1. To determine the subclass of motor neurons generated, the expression profiles of Isl1 and Isl2 were examined. The ectopic motor neurons generated following Hoxa2 misexpression were found to express Isl1 but not Isl2. Therefore, as with Hoxb1, Hoxa2 misexpression also resulted in the generation of motor neurons of branchiomotor subtype identity, as shown by their lateral migration behaviour and by their Isl gene expression patterns. Next, an attempt was made to identify the precise branchiomotor subtypes that are generated after misexpression: the results suggest that the ectopic motor neurons generated following Hoxa2 misexpression are trigeminal-like, while those generated following Hoxb1 misexpression are facial-like. The data demonstrate that at least to a certain extent, and for certain cell types, the singular activities of individual Hox genes (compared to a combinatorial mode of action, for example) are sufficient to impose on neuronal precursor cells the competence to generate distinctly specified cell types. Moreover, since these particular motor neuron subtypes are normally generated in the most anterior domains of Hoxa2 and Hoxb1 expression, respectively, the data support the idea that the main site of individual Hox gene action is in the anteriormost subdomain of their expression, consistent with the phenomenon of posterior dominance (Jungbluth, 1999).

LIM homeodomain codes regulate the development of many cell types, though it is poorly understood how these factors control gene expression in a cell-specific manner. Lhx3 is involved in the generation of two adjacent, but distinct, cell types for locomotion: motor neurons and V2 interneurons. Using in vivo function and protein interaction assays, it has been found that Lhx3 binds directly to the LIM cofactor NLI to trigger V2 interneuron differentiation. In motor neurons, however, Isl1 is available to compete for binding to NLI, displacing Lhx3 to a high-affinity binding site on the C-terminal region of Isl1 and thereby transforming Lhx3 from an interneuron-promoting factor to a motor neuron-promoting factor. This switching mechanism enables specific LIM complexes to form in each cell type and ensures that neuronal fates are tightly segregated (Thaler, 2002).

These studies have defined a biochemical rationale for the early actions of two LIM homeodomain proteins, Isl1 and Lhx3, that have well-established functions in spinal neuron differentiation. In the context of V2 INs, Lhx3 interacts with the LIM-bridging molecule NLI to form tetrameric complexes for the specification of these neurons. The 2NLI:2Lhx3 complex in V2 cells represents the canonical architecture for LIM homeodomain factor complexes based on biochemical and genetic studies in Drosophila. In MN differentiation, the potential to form LIM complexes is complicated by the expression of both Isl1 and Lhx3. These studies indicate that it is not the composite activities of multiple hetero- and homo-meric LIM tetramers, but rather the single action of a novel NLI-mediated hexameric LIM complex that drives MN specification (Thaler, 2002).

Although the assays used in these experiments centered on examining cell specification, it is implicit that the V2 complex and MN complex act differently by regulating distinct genes. This is predicted to occur through the ability of MN-hexamers to recognize different DNA elements from V2-tetramers. The homeodomain of Lhx3 is required for the function of both the MN complex and V2 complex, and thus the specific architecture of the complexes must contribute to the way that Lhx3 is converted from activating V2 IN genes to MN genes in the presence of Isl1 and NLI (Thaler, 2002).

What role does NLI play in the LIM complexes? These studies indicate that the dimerization of NLI is necessary for the proper function of the complexes involved in both IN and MN specification. It remains unclear whether NLI dimerization is necessary for the LIM factors to bind DNA, to activate transcription, or to interact with other factors. In addition to the LIM genes, NLI has been reported to interact with a variety of other transcription factors. In Drosophila, lower-affinity non-LIM interactions have been found to occur in a region of Chip located between residues 439 and 456. The homologous region was included with the amino-terminal region of NLI containing the dimerization domain when creating the chimeric NLI-LIM proteins for this analyses. Therefore, it is possible that in addition to dimerization, NLI also serves as a docking site for additional cofactors involved in neuronal specification. Nevertheless, the NLI:LIM complexes described in this study are able to specify particular neurons at all dorsal-ventral locations, so any important cofactors for the LIM complexes must be present throughout the neural tube and cannot account for their specific functions (Thaler, 2002).

Given the dimerization properties of the LIM factors and NLI, any cell expressing two or more of these proteins is confronted with the possibility of assembling a multitude of complexes. A priori, different cells expressing some of the same LIM factors are expected to assemble overlapping arrays of related transcriptional complexes, thereby creating the potential for activation of inappropriate genes. Not surprisingly, in the developing spinal cord, where numerous transcriptional codes operate to control the acquisition of cell identities, many studies of genetic mutants have detected examples of hybrid cell fates. Clearly the normal mechanisms that control gene expression in the developing spinal cord are designed to restrict inappropriate genes from becoming expressed in order to establish proper cell identity (Thaler, 2002).

Transcriptional synergy between overlapping combinations of transcription factors appears to be one general mechanism for achieving cell type-specific gene regulation. This study has described a mechanism for generating specific transcriptional responses from overlapping LIM-homeodomain transcription factor codes—through the competitive formation of neuronal subtype-specific transcription complexes. Thus, the activity of Lhx3 is regulated by forming different types of complexes in each cellular environment that select different DNA targets to activate. This contrasts with the synergistic mechanism for context-dependent gene activation, where transcription factors are thought to interact with many targets but only activate the subset of genes with the appropriate ensemble of factors present (Thaler, 2002).

The specificity mechanism described in this study raises the question of how cell type-specific LIM complexes might be generated. Several features of Lhx3 and Isl1 appear to be responsible. (1) Lhx3 can interact with both NLI and Isl1; (2) Isl1 directly gates the activity of Lhx3, suppressing its IN activity and activating its MN function; (3) a bias in the formation of LIM complexes in MNs is predicted due to the displacement of Lhx3 from NLI by Isl1 and the presence of a high-affinity binding site for Lhx3 on the C-terminal end of Isl1. Therefore, the formation of hexamers depletes Lhx3 from IN-tetramers. Likewise, the direct binding of Lhx3 to Isl1 reduces the number of 2NLI:2Isl1 tetramers in MNs. Whether 2NLI:2Isl1 complexes have the potential to regulate inappropriate genes in somatic MNs remains unknown. However, hindbrain visceral MNs, dorsal root ganglion sensory neurons, and forebrain neurons express Isl1 in the absence of Lhx3, suggesting 2NLI:2Isl1 tetramers, may be involved in gene regulation within these neurons and therefore may represent an undesirable LIM complex in somatic MNs. The use of multiple interactions involving competition for the same binding site, in this case the LIM domains of Lhx3, represents an additional strategy for using factors in multiple cellular contexts (Thaler, 2002).

Another mechanism that is known to contribute to the clean switch in Lhx3's activity is the involvement of feedback regulatory interactions in MNs by HB9, whose expression is triggered by the MN-hexamer. HB9 probably contributes to the silencing of V2 IN genes in MNs using at least two mechanisms. Studies of mouse mutants indicate that HB9 is necessary for maintenance of high levels of Isl1 in MNs needed to compete with Lhx3 for binding in V2 IN complexes. In addition, HB9 appears to function as a transcriptional repressor that can silence V2 IN genes. Therefore, the initial bias in hexamer formation appears to be sufficient to initiate a cascade of gene expression involving HB9 that serves to further refine the appropriate pattern of gene expression (Thaler, 2002).

Twelve LIM homeodomain genes and four LIM-only genes have been identified in higher vertebrates to date, and many have been implicated in the development of both neuronal and nonneuronal cell types. Within the spinal cord and hindbrain, a number of striking examples of combinatorial LIM codes have been found to label discrete cell populations. Recent studies have also indicated that LIM codes are involved in the specification of particular cell types in the forebrain region of the CNS. Like the example of Isl1 and Lhx3 in this report, cells using multiple LIM factors are confronted with the possibility of assembling a number of LIM complexes with NLI. In MN differentiation this apparent paradox is resolved through the use of multiple protein-protein interactions mediated by the specific LIM domains of Lhx3. The number of cell types and combinatorial codes in which the LIM factors have been implicated suggests these factors may participate in many more cell-specific protein-protein interactions than have currently been identified. In the case of Lhx3, which is also involved in pituitary development, an additional interaction has been found with SLB in these cells. The LIM factors may be particularly well suited to act in combinatorial codes, because LIM domains represent a robust module for mediating numerous protein-protein interactions (Thaler, 2002).

These studies also provide insight into the transcriptional cascades involved in MN and IN development. The LIM homeodomain complexes involved in V2 IN and MN specification are capable of overriding extrinsic signals and progenitor genetic programs in the dorsal neural tube. The relatively late expression and sufficiency of Lhx3/Isl1 codes in directing cell fate decisions provides evidence that combinatorially expressed LIM homeodomain factors execute the progenitor cell repressor programs. This model, however, suggests that MN identity remains undefined beyond the final cell division, since Isl1 is expressed exclusively by postmitotic cells. Similarly, V0 IN specification requires the postmitotic protein, Evx1. The identification of relatively late-acting factors with the capacity to override the initiated preprograms in embryonic progenitor cells may provide the means to generate specific classes of MNs and INs from neural stem cells (Thaler, 2002).

Inductive signaling leads to the coactivation of regulatory pathways for specifying general neuronal traits in parallel with instructions for neuronal subtype specification. Nevertheless, the mechanisms that ensure that these pathways are synchronized have not been defined. To address this, how bHLH proteins Ngn2 and NeuroM controlling neurogenesis functionally converge with LIM-homeodomain (LIM-HD) factors Isl1 and Lhx3 involved in motor neuron subtype specification was investigated. Ngn2 and NeuroM transcriptionally synergize with Isl1 and Lhx3 to specify motor neurons in the embryonic spinal cord and in P19 stem cells. The mechanism underlying this cooperativity is based on interactions that directly couple the activity of the bHLH and LIM-HD proteins, mediated by the adaptor protein NLI. This functional link acts to synchronize neuronal subtype specification with neurogenesis (Lee, 2003).

The underlying transcriptional mechanisms that establish the proper spatial and temporal pattern of gene expression required for specifying neuronal fate are poorly defined. This study characterizes how the Hb9 gene is expressed in developing motoneurons in order to understand how transcription is directed to specific cells within the developing CNS. Non-specific general-activator proteins such as E2F and Sp1 are capable of driving widespread low level transcription of Hb9 in many cell types throughout the neural tube; however, their activity is modulated by specific repressor and activator complexes. The general-activators of Hb9 are suppressed from triggering inappropriate transcription by repressor proteins Irx3 and Nkx2.2. High level motoneuron expression is achieved by assembling an enhancesome on a compact evolutionarily-conserved segment of Hb9 located from -7096 to -6896. The ensemble of LIM-HD and bHLH proteins that interact with this enhancer change as motoneuron development progresses, facilitating both the activation and maintenance of Hb9 expression in developing and mature motoneurons. These findings provide direct support for the derepression model of gene regulation and cell fate specification in the neural tube, as well as establishing a role for enhancers in targeting gene expression to a single neuronal subtype in the spinal cord (Lee, 2004).

Developing motoneurons sequentially express several bHLH proteins, including Ngn2 in the progenitor cells followed by NeuroM in the early postmitotic motoneurons and NeuroD in the more mature cells. Ngn2 and NeuroM have been shown to contribute to the activation of Hb9 during the initial stages of motoneuron development, but it remained unclear whether NeuroD in the mature cells could also stimulate Hb9 expression. To compare the activity of these transcription factors, P19 cells were transfected with expression constructs encoding bHLH proteins together with a luciferase reporter containing seven E box elements. Under these conditions Ngn2 activated the reporter much more than either NeuroM or NeuroD. Despite this inherent difference in transactivation, Ngn2, NeuroM, and NeuroD each synergized in a similar way with the LIM factors Isl1 and Lhx3 to trigger Hb9 expression. Likewise, each bHLH factor dimerizes with E47 and binds to the M50 and M100 E box elements in a sequence-specific manner, and exhibits a similar ability to promote motoneuron differentiation from transfected P19 embryonic carcinoma cells when expressed with Isl1 and Lhx3. Taken together, these findings suggest that the initial activation of Hb9 expression is dependent on Ngn2 and NeuroM as motoneurons become postmitotic, and that NeuroD contributes to the maintenance of Hb9 expression in mature motoneurons (Lee, 2004).

Genetic studies have shown that Hb9 feeds back negatively to modulate its own expression. Whether Hb9 could suppress the activity of its enhancer when LIM and bHLH factors synergize to activate transcription was tested. The native Hb9 protein and the EnR-Hb9 repressor (Hb9 homeodomain linked to eh1 engrailed repressor domain) both inhibited transcription under these conditions, whereas the Hb9-HD and a fusion of Hb9 to the VP16 activation domain (VP16-Hb9) lacked this activity. Thus, in developing motoneurons where Hb9 transcription is synergistically activated, co-repressors such as those recruited by the engrailed fusion (EnR) appear to be involved in negative feedback regulation. Consistent with these findings, Hb9 protein binds in a sequence-specific manner to the ATTA motifs in the enhancer (Lee, 2004).

Onecut transcription factors act upstream of Isl1 to regulate spinal motoneuron diversification

During development, spinal motoneurons (MNs) diversify into a variety of subtypes that are specifically dedicated to the motor control of particular sets of skeletal muscles or visceral organs. MN diversification depends on the coordinated action of several transcriptional regulators including the LIM-HD factor Isl1, which is crucial for MN survival and fate determination. However, how these regulators cooperate to establish each MN subtype remains poorly understood. Using phenotypic analyses of single or compound mutant mouse embryos combined with gain-of-function experiments in chick embryonic spinal cord, this study demonstrates that the transcriptional activators of the Onecut family critically regulate MN subtype diversification during spinal cord development. Evidence that Onecut factors directly stimulate Isl1 expression in specific MN subtypes and are therefore required to maintain Isl1 production at the time of MN diversification. In the absence of Onecut factors, major alterations are observed in MN fate decision characterized by the conversion of somatic to visceral MNs at the thoracic levels of the spinal cord and of medial to lateral MNs in the motor columns that innervate the limbs. Furthermore, Sip1 (Zeb2) was identified as a novel developmental regulator of visceral MN differentiation. Taken together, these data elucidate a comprehensive model wherein Onecut factors control multiple aspects of MN subtype diversification. They also shed light on the late roles of Isl1 in MN fate decision (Roy, 2012).

Saltatory remodeling of Hox chromatin in response to rostrocaudal patterning signals

Efficient transcriptional programming promises to open new frontiers in regenerative medicine. However, mechanisms by which programming factors transform cell fate are unknown, preventing more rational selection of factors to generate desirable cell types. Three transcription factors, Ngn2, Isl1 and Lhx3, are sufficient to program rapidly and efficiently spinal motor neuron identity when expressed in differentiating mouse embryonic stem cells (ESCs). Replacement of Lhx3 by Phox2a leads to specification of cranial, rather than spinal, motor neurons. CHIP analysis of Isl1, Lhx3 and Phox2a binding sites revealed that the two cell fates are programmed by the recruitment of Isl1-Lhx3 and Isl1-Phox2a complexes to distinct genomic locations characterized by a unique grammar of homeodomain binding motifs. These findings suggest that synergistic interactions among transcription factors determine the specificity of their recruitment to cell type-specific binding sites and illustrate how a single transcription factor can be repurposed to program different cell types (Mazzoni, 2013a).

This study exploited the differentiation potential of pluripotent ESCs to study how transcription factor modules control specification of distinct neuronal cell types. Inducible expression of two programming modules differing in one transcription factor led to a rapid and efficient specification of cells expressing key molecular and functional properties of spinal and cranial motor neurons. Isl1 transcription factor changed its genome binding preference when expressed alone or in the context of either 1) the combined expression of Isl1 and Lhx3, together with the proneural gene Ngn2 (NIL factors) or 2) a module in which Lhx3 is replaced by the cranial motor neuron determinant Phox2a (the NIP programming module). Because the factors were expressed in identical cellular context, the different binding preference of Isl cannot be attributed to differential chromatin accessibility or initial presence of distinct cofactors. The data support a model in which Isl forms transcriptional complexes with Lhx3 or Phox2a. The complexes are recruited to condition-specific enhancers with differential motif grammar leading to activation of cell type-specific expression programs and to specification of spinal or cranial motor neurons. These findings have broader consequences for the rational design of programming modules, as mapping an individual transcription factor's DNA binding preference is insufficient to predict its binding and its potential for cellular programming when it is coexpressed with other cooperating programming factors. Systematic computational and experimental analysis of interactions among programming factors, along with decoding the grammar of their cooperative binding motifs, will be a fundamental step toward rational design of programming modules for predictable production of diverse cell types of interest (Mazzoni, 2013a).

The synergistic nature of the programming module's activity could explain why collections of factors are typically required to program terminal cell fate. It is of interest that Oct4, Klf4 and Sox2 (core module) co-occupy regulatory elements in ESCs, suggesting that combinatorial programming modules may be a general developmental strategy. A second set of transcription factors (Myc module) appears to operate in parallel to the core module in pluripotent stem cells. It is therefore anticipated that additional transcriptional modules besides NIL and NIP will contribute to the establishment of terminal motor neuron expression profiles. Notably, the NIL programming module does not activate expression of Hox transcription factors that control specification of motor neuron subtype identity. This is consistent with the recent demonstration that rostro-caudal patterning signals specify motor neuron positional identity by remodeling Hox chromatin landscape during early neural progenitor stages that are bypassed during direct programming by NIL factors (Mazzoni, 2013b). Thus, generic motor neuron identity can be experimentally uncoupled from the Hox-driven program controlling subtype-specific motor neuron properties. Evolution of a generic motor neuron program that operates in parallel with transcription factors controlling subtype-specific programs would provide a versatile and efficient system for diversification of generic motor neurons into distinct subtypes necessary for the assembly of a functioning motor system (Mazzoni, 2013a).

Currently, the identification of effective programming modules relies on empirical testing of combinations of transcription factors expressed in the target cell type. In contrast, the most effective programming module for specification of motor neuron identity is composed of transcription factors expressed only transiently during the transition from motor neuron progenitor to postmitotic state. It is proposed that selection of effective programming modules for other types of nerve cells should focus on transcription factors expressed during similar developmental windows. Without doubt, direct programming of cellular identity will have a substantial effect on human stem cell applications. Differentiation of human pluripotent stem cells to neurons is currently relatively inefficient and slow, taking weeks to months of in vitro culture. Understanding the logic and function of programming modules might not only inform ways to generate cell types refractory to efficient programming by extrinsic patterning signals, but might also substantially accelerate production of homogenous cell populations necessary for human disease modeling, cell-based drug screening and transplantation therapy (Mazzoni, 2013a).

Islet1-mediated activation of the β-catenin pathway is necessary for hindlimb initiation in mice

The transcriptional basis of vertebrate limb initiation, which is a well-studied system for the initiation of organogenesis, remains elusive. Specifically, involvement of the β-catenin pathway in limb initiation, as well as its role in hindlimb-specific transcriptional regulation, are under debate. This study shows that the β-catenin pathway is active in the limb-forming area in mouse embryos. Furthermore, conditional inactivation of β-catenin as well as Islet1, a hindlimb-specific factor, in the lateral plate mesoderm results in a failure to induce hindlimb outgrowth. It was further shown that Islet1 is required for the nuclear accumulation of β-catenin and hence for activation of the β-catenin pathway, and that the β-catenin pathway maintains Islet1 expression. These two factors influence each other and function upstream of active proliferation of hindlimb progenitors in the lateral plate mesoderm and the expression of a common factor, Fgf10. These data demonstrate that Islet1 and β-catenin regulate outgrowth and Fgf10-Fgf8 feedback loop formation during vertebrate hindlimb initiation. This study identifies Islet1 as a hindlimb-specific transcriptional regulator of initiation, and clarifies the controversy regarding the requirement of β-catenin for limb initiation (Kawakami, 2011).

Magnitude of binocular vision controlled by islet-2 repression of a genetic program that specifies laterality of retinal axon pathfinding

Pathfinding of retinal ganglion cell (RGC) axons at the midline optic chiasm determine whether RGCs project to ipsilateral or contralateral brain visual centers, critical for binocular vision. Using Isl2tau-lacZ knockin mice, it has been shown that the LIM-homeodomain transcription factor Isl2 marks only contralaterally projecting RGCs. The transcription factor Zic2 and guidance receptor EphB1, required by RGCs to project ipsilaterally, colocalize in RGCs distinct from Isl2 RGCs in the ventral-temporal crescent (VTC), the source of ipsilateral projections. Isl2 knockout mice have an increased ipsilateral projection originating from significantly more RGCs limited to the VTC. Isl2 knockouts also have increased Zic2 and EphB1 expression and significantly more Zic2 RGCs in the VTC. It is concluded that Isl2 specifies RGC laterality by repressing an ipsilateral pathfinding program unique to VTC RGCs and involving Zic2 and EphB1. This genetic hierarchy controls binocular vision by regulating the magnitude and source of ipsilateral projections and reveals unique retinal domains (Pak, 2004).

These findings indicate that Isl2 normally represses Zic2 expression in RGCs in the VTC and either directly represses EphB1 expression or indirectly through repression of Zic2 and that the increased ipsilateral projection in Isl2-null mice is due to a loss of this repression and upregulation of Zic2 and EphB1. This model is consistent with several pieces of data. (1) Regarding the timing of expression of Isl2 and Zic2 in VTC RGCs, the onset of Isl2 expression in VTC RGCs is similar to that of Zic2: weak Isl2 expression is detected in the VTC as early as E13.5, and moderate levels of Isl2 expression are evident by E14.5, the age when Zic2 expression in VTC RGCs is first detected. (2) Zic2 and EphB1 colocalize in a subset of RGCs distinct from Isl2 RGCs. (3) Increased expression of Zic2 and EphB1 and a significant increase in Zic2-positive RGCs are found in the VTC of Isl2-null retina. (4) The laterality phenotype of Isl2-null mice complements that of Zic2kd/kd and EphB1 mutants (Pak, 2004).

PlexinA4 is necessary as a downstream target of Islet2 to mediate Slit signaling for promotion of sensory axon branching

Slit is a secreted protein known to repulse the growth cones of commissural neurons. By contrast, Slit also promotes elongation and branching of axons of sensory neurons. The reason why different neurons respond to Slit in different ways is largely unknown. Islet2 is a LIM/homeodomain-type transcription factor that specifically regulates elongation and branching of the peripheral axons of the primary sensory neurons in zebrafish embryos. PlexinA4, a transmembrane protein known to be a co-receptor for class III semaphorins, was shown to act downstream of Islet2 to promote branching of the peripheral axons of the primary sensory neurons. Intriguingly, repression of PlexinA4 function by injection of the antisense morpholino oligonucleotide specific to PlexinA4 or by overexpression of the dominant-negative variant of PlexinA4 counteracts the effects of overexpression of Slit2 to induce branching of the peripheral axons of the primary sensory neurons in zebrafish embryos, suggesting involvement of PlexinA4 in the Slit signaling cascades for promotion of axonal branching of the sensory neurons. Colocalized expression of Robo, a receptor for Slit2, and PlexinA4 is observed not only in the primary sensory neurons of zebrafish embryos but also in the dendrites of the pyramidal neurons of the cortex of the mammals, and may be important for promoting the branching of either axons or dendrites in response to Slit, as opposed to the growth cone collapse (Miyashita, 2004).

Isl1 targets in heart development

The vertebrate heart forms initially as a linear tube derived from a primary heart field in the lateral mesoderm. Recent studies in mouse and chick have demonstrated that the outflow tract and right ventricle originate from a separate source of mesoderm that is anterior to the primary heart field. The discovery of this anterior, or secondary, heart field has led to a greater understanding of the morphogenetic events involved in heart formation; however, many of the underlying molecular events controlling these processes remain to be determined. The MADS domain transcription factor MEF2C is required for proper formation of the cardiac outflow tract and right ventricle, suggesting a key role in anterior heart field development. Therefore, as a first step toward identifying the transcriptional pathways upstream of MEF2C, a lacZ reporter gene was introduced into a bacterial artificial chromosome (BAC) encompassing the murine Mef2c locus and this recombinant was used to generate transgenic mice. This BAC transgene was sufficient to recapitulate endogenous Mef2c expression, and comparative sequence analyses revealed multiple regions of significant conservation in the noncoding regions of the BAC. One of these conserved noncoding regions represents a transcriptional enhancer that is sufficient to direct expression of lacZ exclusively to the anterior heart field throughout embryonic development. This conserved enhancer contains two consensus GATA binding sites that are efficiently bound by the zinc finger transcription factor GATA4 and are completely required for enhancer function in vivo. This enhancer also contains two perfect consensus sites for the LIM-homeodomain protein ISL1. These elements are specifically bound by ISL1 and are essential for enhancer function in transgenic embryos. Thus, these findings establish Mef2c as the first direct transcriptional target of ISL1 in the anterior heart field and support a model in which GATA factors and ISL1 serve as the earliest transcriptional regulators controlling outflow tract and right ventricle development (Dodou, 2004).

The LIM homeodomain transcription factor Islet1 (Isl1) is expressed in both foregut endoderm and cardiogenic mesoderm and is required for earliest stages of heart development. isl1 is also required upstream of Shh. In isl1 null mice, Sonic hedgehog (Shh) is downregulated in foregut endoderm. Shh signals through the unique activating receptor smoothened (Smo). To investigate the role of hedgehog signaling in the isl1 domain, smo utilizing isl1-cre was ablated. Isl1-cre;smo mutants exhibit cardiovascular defects similar to those observed in Shh null mice, defining a spatial requirement for hedgehog signaling within isl1 expression domains for aortic arch and outflow tract formation. Semaphorin signaling through neuropilin receptors npn1 and npn2 is required for aortic arch and outflow tract formation. Expression of npn2 is downregulated in isl1-cre;smo mutants, suggesting an isl1/Shh/npn pathway required to affect morphogenesis at the anterior pole of the heart (Lin, 2007a).

Recent studies have demonstrated that the LIM homeodomain transcription factor Islet1 (Isl1) marks pluripotent cardiovascular progenitor cells and is required for proliferation, survival, and migration of recently defined second heart field progenitors. Factors that are upstream of Isl1 in cardiovascular progenitors have not yet been defined. This study demonstrates that β-catenin is required for Isl1 expression in cardiac progenitors, directly regulating the Isl1 promoter. Ablation of β-catenin in Isl1-expressing progenitors disrupts multiple aspects of cardiogenesis, resulting in embryonic lethality at E13. β-Catenin is also required upstream of a number of genes required for pharyngeal arch, outflow tract, and/or atrial septal morphogenesis, including Tbx2, Tbx3, Wnt11, Shh, and Pitx2. These findings demonstrate that β-catenin signaling regulates proliferation and survival of cardiac progenitors (Lin, 2007b).

Ssdp1 regulates head morphogenesis of mouse embryos by activating the Lim1-Ldb1 complex

The transcriptional activity of LIM-homeodomain (LIM-HD) proteins is regulated by their interactions with various factors that bind to the LIM domain. Reduced expression of single-stranded DNA-binding protein 1 (Ssdp1), which encodes a co-factor of LIM domain interacting protein 1 (Ldb1), in the mouse mutant headshrinker (hsk) disrupts anterior head development by partially mimicking Lim1 mutants. Although the anterior visceral endoderm and the anterior definitive endoderm, which together comprise the head organizer, were able to form normally in Ssdp1hsk/hsk mutants, development of the prechordal plate was compromised. Head development is partially initiated in Ssdp1hsk/hsk mutants, but neuroectoderm tissue anterior to the midbrain-hindbrain boundary is lost, without a concomitant increase in apoptosis. Cell proliferation is globally reduced in Ssdp1hsk/hsk mutants, and approximately half also exhibit smaller body size, similar to the phenotype observed in Lim1 and Ldb1 mutants. Ssdp1 contains an activation domain and is able to enhance transcriptional activation through a Lim1-Ldb1 complex in transfected cells, and Ssdp1 interacts genetically with Lim1 and Ldb1 in both head development and body growth. These results suggest that Ssdp1 regulates the development of late head organizer tissues and body growth by functioning as an essential activator component of a Lim1 complex through interaction with Ldb1 (Nishioka, 2005).

back to islet Evolutionary Homologs part 1/3 | part 2/3

islet/tailup: Biological Overview | Regulation | Developmental Biology | Effects of Mutation | References

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.