Progenitor cells in the mouse olfactory epithelium generate over a thousand subpopulations of neurons, each expressing a unique odorant receptor (OR)gene. This event is under the control of spatial cues, since neurons in different epithelial regions are restricted to express region-specific subsets of OR genes. We show that progenitors and neurons express the LIM-homeobox gene Lhx2 and that neurons in Lhx2-null mutant embryos do not diversify into subpopulations expressing different OR genes and other region-restricted genes such as Nqo1 and Ncam2. Lhx2-/- embryos have, however, a normal distribution of Mash1-positive and neurogenin 1-positive neuronal progenitors that leave the cell cycle, acquire pan-neuronal traits and form axon bundles. Increased cell death in combination with increased expression of the early differentiation marker Neurod1, as well as reduced expression of late differentiation markers (Gαolf and Omp), suggests that neuronal differentiation in the absence of Lhx2 is primarily inhibited at, or immediate prior to, onset of OR expression. Aberrant regional expression of early and late differentiation markers, taken together with unaltered region-restricted expression of the Msx1 homeobox gene in the progenitor cell layer of Lhx2-/- embryos, shows that Lhx2 function is not required for all aspects of regional specification of progenitors and neurons. Thus,these results indicate that a cell-autonomous function of Lhx2 is required for differentiation of progenitors into a heterogeneous population of individually and regionally specified mature olfactory sensory neurons.

A major challenge in our understanding of how axonal projection maps are generated is to identify gene regulatory pathways that generate neuronal diversity, both with regard to response properties and axon target selection of individual neurons. A notable example is the highly heterogeneous population of olfactory sensory neurons (OSNs) in the olfactory epithelium(OE) of the nose. Orderly afferent projections of approximately 1000 different subpopulations of mouse OSNs, each expressing a unique odorant receptor (OR)gene, form a projection map decoding olfactory sensory information in the olfactory bulb of the brain (Buck and Axel,1991; Ressler et al.,1994; Vassar et al.,1994). The scattered distribution of cell bodies of a given OR-specific subpopulation is confined to one of several distinct regions of OE. OSNs expressing different subsets of OR genes thus partition OE into zones(Ressler et al., 1993; Vassar et al., 1993). Some overlap in OR expression between zones divides OE roughly into a dorsomedial,a middle and a ventrolateral zone (Iwema et al., 2004). The zonal organization of OE is also evident from the differential expression of other types of genes, such as the neural cell adhesion molecule 2 (Ncam2/Ocam/Rncam) and NADPH:quinone oxidoreductase 1 (Nqo1/DT-diaphorase)(Alenius and Bohm, 1997; Gussing and Bohm, 2004; Yoshihara et al., 1997). Nqo1 catalyzes reduction quinones and is co-expressed with OR genes of the dorsomedial zone (Gussing and Bohm,2004). Ncam2, which is co-expressed with ORs of the middle and the ventrolateral zones, has been shown to play a role in axonal guidance of OSNs(Alenius and Bohm, 2003).

OE originates from ectodermally derived neurogenic placodes and the appearance of cell layers and zones becomes evident around embryonic day (E)12.5-13.5 in mouse (Cau et al.,1997; Sullivan et al.,1995). The three major cell layers are: a superficial layer of sustentacular (supporting) cells; a basal cell layer with dividing immediate neuronal progenitors; and an intermediate cell layer containing OSNs. Also progenitor and sustentacular cells differentially express certain genes in a manner that correlates with their zonal position in OE(Miyawaki et al., 1996; Norlin et al., 2001; Tietjen et al., 2003). The significance of zone-specific gene regulation in the progenitor and sustentacular cell layers is not known.

Functional analyses of certain basic helix-loop-helix (bHLH) transcription factors that are expressed in the progenitor cells in all zones have shown that Mash1 (Ascl1 – Mouse Genome Informatics) and neurogenin 1 (Ngn1)are required for the development of OSNs(Cau et al., 2002; Guillemot et al., 1993). Mash1 is required for the survival of OSN progenitors at an early stage in the OSN lineage, whereas the function of Ngn1 appears important in initiating differentiation, presumably via the bHLH protein Neurod1, which is transiently expressed at the onset of differentiation(Cau et al., 2002). Despite these advances in knowledge of the essential roles played by bHLH in ensuring neurogenesis in OE, transcription factors regulating diversification of OSNs,including the choice of a given OR gene or the formation of OE zones, have not been identified. A candidate gene that is expressed in a Mash1-dependent manner in OE is Lhx2 (Cau et al.,2002; Tietjen et al.,2003). Lhx2 is a LIM-homeodomain transcription factor that is implied in the specification of several aspects of the neuronal phenotype(Monuki et al., 2001; Porter et al., 1997; Xu et al., 1993). Analyses of Lhx2-deficient mice have shown that Lhx2 is required for the formation of the optic cup that gives rise to the multilayered neural retina(Porter et al., 1997). In addition, lack of Lhx2 function results in agenesis of the hippocampus and profound losses of cortical progenitors and neurons(Monuki et al., 2001; Porter et al., 1997). Other LIM homeobox family members are required for the diversification of neuronal subtypes in the spinal cord (Shirasaki and Pfaff, 2002). A specific role for Lhx2 in the generation of neuronal subtypes has, however, remained elusive. In the present study we have analyzed the function of Lhx2 during neurogenesis in OE. We provide evidence that the expression of Lhx2 in the OSN lineage is required for the generation of OR-and zone-specific subpopulations of OSNs during differentiation.

Mice

All mice were maintained at the animal facility at Umeå University under pathogen-free conditions. Generation of mice with a targeted Lhx2 gene has been described elsewhere(Porter et al., 1997), and mice and embryos were genotyped as previously described(Monuki et al., 2001). Lhx2+/– mice were maintained on a mixed 129/Sv×C57BL/6 background and were mated to generate embryos for analyses. The morning of the vaginal plug was considered as E0.5. No obvious difference in gene expression patterns could be observed between wild-type or heterozygous E14.5-15.5 embryos and differences in gene expression between E14.5 and E15.5 Lhx2-/- embryos were not observed. All animal experiments were approved by the ethical committee at UmeåUniversity.

In-situ hybridization

Embryos were postfixed in 4% paraformaldehyde (PFA), cryoprotected (30%sucrose in PBS), embedded in Tissue-Tek OCT compound and cryosectioned. Cryosections (10 μm) for in-situ hybridization using 35S-labeled probes were pretreated according to Breitschopf et al.(Breitschopf et al., 1992) and hybridization and washing conditions were as described(Sassoon and Rosenthal, 1993). Slides were dehydrated and processed for autoradiography using NTB-emulsion(Kodak), and exposure was for 5-14 days at 4°C. In-situ hybridization using digoxygenin (Dig)-labeled probes was performed as described (Schaeren-Wiemers and Gerfin-Moser, 1993), with some modifications. Briefly sections were treated with 5 μg/ml proteinase K (Roche) in PBS for 15 minutes in Fast Red (Roche). Probes specific for Gαolf, Omp and OR genes (M40, K20,L45, A16, and M50) have been described previously(Berghard et al., 1996; Ressler et al., 1993; Sullivan et al., 1995). The Ncam2-specific probe corresponded to the extracellular Ncam2domain (Alenius and Bohm,1997). The Lhx2 probe (460-1750 bp, accession no. NM_010710)hybridized to both wild-type and mutated Lhx2 transcripts(Monuki et al., 2001). Probes corresponded to published sequences of Nqo1 (94-1626 bp, accession no. NM_008706), Msx1 (894-1600 bp, accession no. BC016426), and Alk6 (Bmpr1b – Mouse Genome Informatics) (1-1935 bp,accession no. MMALK6A). The Mash1-, Ngn1- and Neurod1-specific probes corresponded to the published sequences of Mash1 (256-900 bp, accession no. MUSHASH1X), Ngn1 (75-840 bp, accession no. YO9166), and Neurod1 (1-1207 bp, accession no. BC018241). Hoechst 33258 (Sigma) was used as nuclear counterstain.

Immunohistochemistry

Embryos were fixed in 4% PFA for 1-2 hours at room temperature and cryoprotected at 4°C for 24 hours. Retrieval of Nquo1, Rncam and Gap43 antigens was achieved by boiling in 10 mM citrate buffer (pH 6.0) for 10 minutes. Sections were then incubated for 1 hour in blocking solution (3%normal donkey serum, 0.3% Triton X-100 in PBS) followed by overnight incubation at 4°C in blocking solution containing affinity-purified anti-Ncam2 (1:100 dilution (Alenius and Bohm, 2003), anti-Gap43 (1:1000 dilution, AB5220, Lot#22010885,Chemicon), anti-caspase-3 antibody (1:1000 dilution, 557035, Lot#M051497,PharMingen), anti-Omp (1:2000 dilution, gift from Dr Margolis), anti-Ncam(1:100 dilution, 556323, clone 12F11, PharMingen), phospho-H3 (1:500 dilution,06-570, Lot#24019, Upstate biotech), anti-Stathmin/SCG10 antibody (1:1000 dilution (Holmfeldt et al.,2003), and neuronal class III β-tubulin (1:2000 dilution,PRB-435P, Nordic BioSite AB), anti-LH2A [1:500(Liem et al., 1997)]. Positive immunoreaction was visualized after a 1 hour incubation with Alexa Fluor 546 donkey anti-goat IgG (A-11056) or Alexa Fluor 488 donkey anti-rabbit IgG(A-21206, Molecular Probes) diluted 1:250 in blocking solution. Sections were counterstained with Hoechst. Microphotographs of immunohistochemistry and in-situ hybridization analyses were taken using light, fluorescent or dark-field optics on a Zeiss Axioskop microscope with a Hamamatsu digital CCD camera. Confocal microscopy was performed on a Nikon confocal microscope. Images were processed using Adobe Photoshop 7.0.1. Brightness and contrast adjustments of images were done linearly.

Lack of OR gene expression in Lhx2-null embryos

Double Lhx2 and neuronal class III β-tubulin (TubIII)immunohistochemistry analysis of E15.5 mouse embryos showed nuclear Lhx2 in TubIII-positive OSNs (Fig. 1A-C). A fraction of TubIII-negative cells in the basal progenitor cell layer stained positive with the Lhx2 antibody (arrow in 1C). The sustentacular cell layer or in cells located in lamina propria under the OE were negative. OR gene expression was first observed at E11.5 in the developing OE. The number of OR-positive cells increases dramatically thereafter and zonal OR expression is apparent by E13(Sullivan et al., 1995). Homozygous Lhx2 knockout mouse (Lhx2-/-) die in utero after E15.5 (Porter et al.,1997). Lhx2-/- and littermate control embryos(Lhx2+/– and Lhx2+/+) at two ages (E14.5 and E15.5) were analyzed in this study. Lhx2 and OR gene expression was examined by in-situ hybridization(Fig. 1D-J). Lhx2 mRNA was, in concordance with the result of the immunohistochemical analysis,expressed in both OSNs and a fraction of cells in the basal progenitor cell layer in control embryos (Fig. 1D). The targeted Lhx2 allele produces a fusion transcript of Lhx2 exon 1 and the neomycin resistance gene that allows for the identification of cells that have an active Lhx2promoter also in Lhx2-/- embryos(Monuki et al., 2001). Importantly, we detected a similar cell layer organization and distribution of Lhx2-promoter active cells in control and homozygous mutant mice(Fig. 1D,E). The result suggested that Lhx2 is not required for formation of the progenitor and neuronal cell layers of OE. Next we analyzed for the expression of different OR genes that are expressed in different zones, i.e. K21(Olfr145 – Mouse Genome Informatics), L45, M50(Olfr6 – Mouse Genome Informatics) and A16 (Olfr140– Mouse Genome Informatics). As expected, all four OR probes hybridized to scattered OSNs in sections from control mice(Fig. 1F-I). Interestingly,parallel analyses of Lhx2-/- littermates revealed no hybridizing cells (Fig. 1F-I),even though sections throughout the anterior–posterior extent of the nasal cavity were examined. Neither did in-situ hybridization analyses using more sensitive radioactively labeled cRNA probes reveal any OR positive OSNs in mutant embryos (Fig. 1J). These results indicated that the function of Lhx2 was required for OR expression.

Fig. 1.

Unaltered distribution of Lhx2 expression and lack of OR expression in OE of Lhx2 -/- embryos. (A-C)Double-immunohistochemical analysis of Lhx2 and TubIII expression in OE of E15.5 mouse embryo is shown. (A) Nuclear Lhx2 immunoreactivity in the neuronal cell layer (N) and in a fraction of cells in the basal progenitor cell layer(P). No Lhx2-positive cells are present in the sustentacular cell layer (S)and the lamina propria (l.p.). (B) TubIII immunoreactivity in neuronal soma and neuronal processes. (C) Images in A and B combined, showing that Lhx2 and TubIII are co-expressed in OSNs. A limited number of TubIII-negative Lhx2-positive cells are present in the progenitor cell layer (arrows in C).(D-I) In-situ hybridization analyses of serial OE sections from control (left panel) and Lhx2-/- embryos (right panel) with Dig-labeled cRNA probes that hybridize to Lhx2 and mutated Lhx2transcripts (Lhx2m) (D,E), and OR genes (K21, L45,M50 and A16) (F-I). (J) Autoradiographs of in-situ hybridization analyses with a 35S-labeled cRNA probe specific to the OR gene K20. Scattered cells expressing the different OR genes are present in control mice. The zones in which each the OR gene is expressed is indicated,i.e. dorsomedial (dm), the middle (m) or the ventrolateral (vl) zone. In Lhx2 -/- mice, there are no cells expressing ORs in any epithelial region (representative results are shown).

Fig. 1.

Unaltered distribution of Lhx2 expression and lack of OR expression in OE of Lhx2 -/- embryos. (A-C)Double-immunohistochemical analysis of Lhx2 and TubIII expression in OE of E15.5 mouse embryo is shown. (A) Nuclear Lhx2 immunoreactivity in the neuronal cell layer (N) and in a fraction of cells in the basal progenitor cell layer(P). No Lhx2-positive cells are present in the sustentacular cell layer (S)and the lamina propria (l.p.). (B) TubIII immunoreactivity in neuronal soma and neuronal processes. (C) Images in A and B combined, showing that Lhx2 and TubIII are co-expressed in OSNs. A limited number of TubIII-negative Lhx2-positive cells are present in the progenitor cell layer (arrows in C).(D-I) In-situ hybridization analyses of serial OE sections from control (left panel) and Lhx2-/- embryos (right panel) with Dig-labeled cRNA probes that hybridize to Lhx2 and mutated Lhx2transcripts (Lhx2m) (D,E), and OR genes (K21, L45,M50 and A16) (F-I). (J) Autoradiographs of in-situ hybridization analyses with a 35S-labeled cRNA probe specific to the OR gene K20. Scattered cells expressing the different OR genes are present in control mice. The zones in which each the OR gene is expressed is indicated,i.e. dorsomedial (dm), the middle (m) or the ventrolateral (vl) zone. In Lhx2 -/- mice, there are no cells expressing ORs in any epithelial region (representative results are shown).

Lack of zone-specific Nqo1 and Ncam2 in Lhx2-null embryos

Lack of OR expression in Lhx2-/- embryos suggested that the function of Lhx2 was important specifically for OR expression. However, since mutant mice lacked expression of several ORs that are expressed in different zones, an alternative interpretation was that Lhx2 is required for zone-specific gene expression in general and not OR gene expression per se. To address this possibility, we analyzed for the expression of other zone-specific markers. Nqo1 is co-expressed with ORs of the dorsomedial zone, whereas Ncam2 is co-expressed with ORs of both the middle and the ventrolateral zones (Alenius and Bohm, 2003; Gussing and Bohm, 2004). In-situ hybridization analyses revealed zonal Nqo1 and Ncam2 expression in control mice(Fig. 2A, parts B,C and Fig. 2B, parts A,B), whereas no hybridizing cells were observed in the OE of Lhx2-/-embryos (Fig. 2A, parts E,F and Fig. 2B, parts C,D). Other cell types that hybridized to Nqo1 and Ncam2 probes were, however, present in both control and Lhx2-/- embryos (arrowheads in Fig. 2A, parts B,E,F). Thus,the function of Lhx2 was required for the development of OSNs that express both ORs and zone-specific genes.

Fig. 2.

No expression of the zone-specific genes Nqo1 and Ncam2in OE of Lhx2-/- embryos. (A) Autoradiographs of in-situ hybridization analyses of consecutive coronal sections (dorsal up and medial left) of one nasal cavity of control embryos (A-C) and Lhx2-/- embryos (D-F). Sections were hybridized with 35S-labeled cRNA probes specific for Lhx2 and mutated Lhx2 transcripts (Lhx2m) (A,D), Nqo1(B,E) and Ncam2 (C,F). Lhx2 expression is distributed throughout the OE in both control (A) and Lhx2-/- mice(D). Nqo1 is expressed in OSNs located in the dorsomedial zone in control embryos (arrow in B), whereas no signal is present in the neuronal cell layer of Lhx2-/- embryos (E). Ncam2 is expressed in OSNs located in the middle and the ventrolateral zones in control embryos (arrow in C), whereas no hybridizing signal is present in the neuronal cell layer of Lhx2-/- embryos (F). The border between neurons in the dorsomedial zone and the middle zone is indicated (dotted line in B,C). Nqo1-specific hybridization to cells in lamina propria(arrowheads in B,E) and Ncam2-specific hybridization to cells in the telencephalon (arrowhead in F) are also indicated. (B) High-power magnifications of OE sections hybridized with Dig-labeled cRNA probes specific for Nqo1 (A,C) and Ncam2 (B,D) are shown. Both probes hybridize to OSN in control embryos (A,B), whereas no signal over background is evident in sections of Lhx2-/- embryos (C,D).

Fig. 2.

No expression of the zone-specific genes Nqo1 and Ncam2in OE of Lhx2-/- embryos. (A) Autoradiographs of in-situ hybridization analyses of consecutive coronal sections (dorsal up and medial left) of one nasal cavity of control embryos (A-C) and Lhx2-/- embryos (D-F). Sections were hybridized with 35S-labeled cRNA probes specific for Lhx2 and mutated Lhx2 transcripts (Lhx2m) (A,D), Nqo1(B,E) and Ncam2 (C,F). Lhx2 expression is distributed throughout the OE in both control (A) and Lhx2-/- mice(D). Nqo1 is expressed in OSNs located in the dorsomedial zone in control embryos (arrow in B), whereas no signal is present in the neuronal cell layer of Lhx2-/- embryos (E). Ncam2 is expressed in OSNs located in the middle and the ventrolateral zones in control embryos (arrow in C), whereas no hybridizing signal is present in the neuronal cell layer of Lhx2-/- embryos (F). The border between neurons in the dorsomedial zone and the middle zone is indicated (dotted line in B,C). Nqo1-specific hybridization to cells in lamina propria(arrowheads in B,E) and Ncam2-specific hybridization to cells in the telencephalon (arrowhead in F) are also indicated. (B) High-power magnifications of OE sections hybridized with Dig-labeled cRNA probes specific for Nqo1 (A,C) and Ncam2 (B,D) are shown. Both probes hybridize to OSN in control embryos (A,B), whereas no signal over background is evident in sections of Lhx2-/- embryos (C,D).

Inhibited OSN differentiation and increased apoptosis in Lhx2-/- embryos

Formally, the phenotype observed might stem from a complete lack of neuronal progenitors, altered commitment of multipotent progenitors or that neuronal differentiation was inhibited at a stage prior to onset of OR and zone-specific gene expression. To address the question of whether Lhx2 is required for the generation of OSN progenitors, we analyzed for the expression of Mash1 and Ngn1. It has been suggested that the developing OE contains two populations of neuronal progenitors: a Mash1-positive population that divides in the apical part of OE and a population of Mash1- and Ngn1-positive secondary progenitor cells that settle on the basal side of OE, where they continue to divide before differentiating into OSNs (Cau et al., 1997). Expression levels and distribution of Mash1- and Ngn1-positive cells in OE of control and Lhx2 mutant mice were similar(Fig. 3A,B). Thus, Lhx2 function did not appear to be required for Mash1/Ngn1-dependent neurogenesis. Ngn1 regulates the differentiation-promoting gene Neurod1, which is transiently expressed during the onset of OSN differentiation(Cau et al., 2002). Interestingly, Lhx2 mutant embryos had an increased number of Neurod1-positive cells (Fig. 3C). Neurod1 was expressed in the one- to two-cell thick layer of progenitor cells in control mice, whereas Neurod1-positive cells had expanded into the presumptive neuronal cell layer in Lhx2 mutants(Fig. 3C). Thus, it appeared that lack of Lhx2 resulted in accumulation of Neurod1-positive progenitor cells and/or immature OSNs. To characterize this phenotype further,we analyzed for expression of OSN-specific genes Gαolf and Omp, which are expressed at terminal stages of OSN differentiation. The gene for the G-protein alpha subunit Gαolf is turned on at a stage between E13.5-16.5, which is at least 2 days subsequent to the onset of OR expression (Sullivan et al.,1995). Gαolf was expressed in neurons of both control and Lhx2-/- embryos at E14.5-15.5(Fig. 3D). This result suggested that the observed phenotype was not a consequence of a developmental delay in the absence of Lhx2. The expression level of Gαolf was, however, reduced compared with littermate control embryos (Fig. 3D). We next analyzed for the expression of Omp, which is a hallmark of mature OSN(Youngentob et al., 2003). Studies of neuronal differentiation in the OE show that ORs are expressed prior to onset of Omp expression(Iwema and Schwob, 2003). In-situ hybridization analysis showed that a limited number of neurons in Lhx2-/- mice expressed Omp to markedly reduced levels(Fig. 3E). The total number of Omp-positive cells was 133±61 and 13±5 (P<0.001,Student's t-test) per section and nasal cavity in control and Lhx2-/- embryos, respectively. The presence of neurons expressing Omp in Lhx2-/- mice was confirmed by immunohistochemical analyses (Fig. 3F). Unexpectedly, analyses of sections throughout the nasal cavity revealed that Omp-positive and Gαolf-positive cells were located selectively in the dorsomedial part of OE (for result and discussion see below). The reduced expression of Omp and Gαolf suggested that neuronal maturation and/or survival was reduced. To address if reduced expression of Gαolf and Omp coincided with increased apoptosis, we utilized activated(cleaved) caspase 3 immunohistochemistry analysis. Caspase 3-positive cells were evenly distributed throughout OE (data not shown) and the number of cells with activated caspase 3 in the progenitor and neuronal layers was 5.4±2.1 and 19.4±8.9 (P<0.001, Student's t-test) per OE section in control and Lhx2-/-embryos, respectively. The aberrant differentiation of OSNs in Lhx2-/- embryos was thus associated with an overall 3.5-fold increased rate of apoptosis. Collectively, these results indicated that a fraction of neurons in the dorsomedial OE acquires an OSN identity in the absence of Lhx2 but fails to become and/or survive as OR- and zonally specified bona fide OSNs.

Fig. 3.

Unaltered neurogenesis and altered neuronal differentiation in Lhx2-/- embryos. In-situ hybridization analyses of OE sections with Dig-labeled cRNA probes specific for Mash1 (A), Ngn1 (B), Neurod1 (C), Gαolf (D), and Omp (E). Control embryos (left panel) and Lhx2-/-embryos (right panel) show an equal distribution and number of apical and basal progenitor cells expressing Mash1 (A) and basal progenitors expressing Ngn1 (B). The OE in Lhx2-/- embryos contains an increased number of progenitors and/or neurons that express Neurod1 (C), whereas the expression of OSN-enriched and late differentiation markers Gαolf (D) and Omp (E) is reduced. (F) Immunohistochemical analyses of the dorsomedial zone (dorsal left and lateral up) with a reduced number of Omp-positive cells in OE of Lhx2-/- embryos.

Fig. 3.

Unaltered neurogenesis and altered neuronal differentiation in Lhx2-/- embryos. In-situ hybridization analyses of OE sections with Dig-labeled cRNA probes specific for Mash1 (A), Ngn1 (B), Neurod1 (C), Gαolf (D), and Omp (E). Control embryos (left panel) and Lhx2-/-embryos (right panel) show an equal distribution and number of apical and basal progenitor cells expressing Mash1 (A) and basal progenitors expressing Ngn1 (B). The OE in Lhx2-/- embryos contains an increased number of progenitors and/or neurons that express Neurod1 (C), whereas the expression of OSN-enriched and late differentiation markers Gαolf (D) and Omp (E) is reduced. (F) Immunohistochemical analyses of the dorsomedial zone (dorsal left and lateral up) with a reduced number of Omp-positive cells in OE of Lhx2-/- embryos.

Differentiation is inhibited at, or just prior to, onset of OR expression

Lhx2 deficiency appeared to influence differentiation primarily at a stage that was subsequent to onset of Neurod1 expression and prior to onset of Gαolf and Omp expression. To determine if there was an increased number of Neurod1-positive progenitors and/or Neurod1-positive immature neurons in Lhx2-/- embryos we analyzed for Neurod1 expression in relation to dividing cells and onset of early pan-neuronal differentiation markers. Immunohistochemical analyses with antibodies specific for Ncam1, Gap43, Stmn/SCG10 (Stmn2 –Mouse Genome Informatics) and TubIII showed that immature neurons were present throughout OE in both control and mutant embryos(Fig. 4A-F,I,J). Moreover, the staining patterns showed immunoreactivity in axon bundles in the lamina propria and in the OE of Lhx2-/- mice (arrows in 4A-F,I,J). Thus, progenitor cells appeared to leave the cell cycle and differentiate to a stage associated with the onset of pan-neuronal markers and axonal outgrowth. Analyses of serial sections revealed that regions in OE that showed high Neurod1 and TubIII expression correlated(Fig. 4I-L). Confocal microscopy indicated an increased number of cells that co-expressed Neurod1 and TubIII in Lhx2-/- embryos(Fig. 4M,N) which is compatible with the suggestion that Lhx2-/- embryos have a larger fraction of newly differentiated Neurod1-positive neurons and not proliferating progenitors. To substantiate this suggestion, we analyzed for the mitosis marker phosphorylated histone 3 (phospho-H3). The increased number of Neurod1-positive cells in Lhx2-/- embryos did not correlate to an increase in dividing progenitors(Fig. 4G,H,O,P). Thus, these results indicated that the transient differentiation stage, at which Neurod1 and pan-neuronal markers are co-expressed, is prolonged in Lhx2-/- embryos. Since there is a lag between onset of pan-neuronal gene expression and onset of OR expression(Iwema and Schwob, 2003), this suggests that neuronal differentiation in Lhx2-/- embryos is primarily inhibited at a stage that is normally associated with the onset of OR gene expression.

Fig. 4.

The neuronal cell layer in OE of Lhx2-/- embryos largely contains newly differentiated neurons that have acquired pan-neuronal traits and form axon bundles. Coronal sections (dorsal up and medial left) of one nasal cavity from control embryos (A,C,E,G,I,K,M,O) and Lhx2-/- embryos (B,D,F,H,J,L,N,P). (A-F)Immunohistochemical analyses for expression of immature neuronal markers is shown. Ncam1 (A,B), Gap43 (c-d), and Stathmin/SCG10 (E,F) are expressed in OE and axon bundles of the lamina propria (arrows) in both control and Lhx2-/-embryos. (G,H) Double phospho-H3 (PH3; in green) immunohistochemistry and Neurod1 (ND; in red) in-situ hybridization analyses showing that the increased expression of Neurod1, primarily apparent in the ventral OE(arrowhead in H), does not correlate with the number and distribution of mitotic cells. (I-L) Analyses of consecutive OE sections, showing that cells immunoreactive using an anti-TubIII antibody (I,J) and cells that express Neurod1 transcripts (K,L) co-localize predominantly in ventrolateral regions of OE (arrowheads in J,L). Expression of TubIII in axon bundles of the lamina propria is indicated (arrows in I,J). (M,N) High magnification confocal images of double TubIII immunohistochemical (in green) and Neurod1 in-situ hybridization (in red) analyses showing an increased number of cells in Lhx2-/- embryos that co-express Neurod1 and TubIII (yellow signal; insert in N). (O,P) High magnification confocal images of double phospho-H3 immunohistochemical (in green) and Neurod1in-situ hybridization (in red) analyses showing that OE in Lhx2-/- embryos does not contain an increased number of progenitors that co-express phospho-H3 and Neurod1 (in yellow;arrowheads).

Fig. 4.

The neuronal cell layer in OE of Lhx2-/- embryos largely contains newly differentiated neurons that have acquired pan-neuronal traits and form axon bundles. Coronal sections (dorsal up and medial left) of one nasal cavity from control embryos (A,C,E,G,I,K,M,O) and Lhx2-/- embryos (B,D,F,H,J,L,N,P). (A-F)Immunohistochemical analyses for expression of immature neuronal markers is shown. Ncam1 (A,B), Gap43 (c-d), and Stathmin/SCG10 (E,F) are expressed in OE and axon bundles of the lamina propria (arrows) in both control and Lhx2-/-embryos. (G,H) Double phospho-H3 (PH3; in green) immunohistochemistry and Neurod1 (ND; in red) in-situ hybridization analyses showing that the increased expression of Neurod1, primarily apparent in the ventral OE(arrowhead in H), does not correlate with the number and distribution of mitotic cells. (I-L) Analyses of consecutive OE sections, showing that cells immunoreactive using an anti-TubIII antibody (I,J) and cells that express Neurod1 transcripts (K,L) co-localize predominantly in ventrolateral regions of OE (arrowheads in J,L). Expression of TubIII in axon bundles of the lamina propria is indicated (arrows in I,J). (M,N) High magnification confocal images of double TubIII immunohistochemical (in green) and Neurod1 in-situ hybridization (in red) analyses showing an increased number of cells in Lhx2-/- embryos that co-express Neurod1 and TubIII (yellow signal; insert in N). (O,P) High magnification confocal images of double phospho-H3 immunohistochemical (in green) and Neurod1in-situ hybridization (in red) analyses showing that OE in Lhx2-/- embryos does not contain an increased number of progenitors that co-express phospho-H3 and Neurod1 (in yellow;arrowheads).

Lack of Lhx2 function reveals zone-specific differences in neuronal maturation

Although neurons in Lhx2-/- embryos were not zonally specified with regard to OR, Nqo1 or Ncam2 expression, we found evidence that Lhx2 deficiency did not abolish all zonal characteristics of the OE. The ventrolateral region contained a higher fraction of Neurod1-positive cells than did the dorsomedial region (arrowheads in Fig. 4H,L). This regional difference in distribution of Neurod1-positive cells was more apparent in Lhx2-/- embryos than controls. Analyses of consecutive sections indicated that the increased expression of Neurod1 in the ventrolateral part coincided with a high expression of TubIII (arrowheads in Fig. 4J,L). Since phospho-H3 immunoreactive cells were distributed equally throughout the OE(Fig. 4G,H), this result suggested that the ventrolateral region contained a larger fraction of newly differentiated postmitotic neurons compared with the dorsomedial OE region in Lhx2-/- embryos. Interestingly, neurons expressing the late differentiation markers (Gαolf and Omp) were selectively located in the dorsomedial OE region in Lhx2-/- embryos (Fig. 5A-D). The Gαolf- and Omp-positive region appeared to correspond to the dorsomedial zone, namely the dorsal meatus, part of the medial septum, and the tips of the ethmoturbinate projecting into the nasal cavity (Fig. 5B,D and compare with expression of Nqo1 in Fig. 5E). To confirm the unequal distribution of Omp expression, we quantified the number of Omp-positive cells that were located on either side of a tentative half circle dividing the OE into two roughly equal parts, one dorsomedial and one ventrolateral part. This analysis revealed that 59.7±14.5% and 95.2±10.9% (P<0.001, Student's t-test) of the Omp-positive cells were located in the dorsomedial part in control and Lhx2-/- embryos, respectively. Thus, the absence of Lhx2 expression revealed zone-specific differences in the regulated control of neuronal differentiation.

Fig. 5.

Abnormal region-specific OSN differentiation and normal region-specific gene expression in sustentacular and progenitor cells. In-situ hybridization analyses of serial coronal (dorsal up and medial left) sections of OE from control (A,C,E,G,I,K,M) and Lhx2-/- embryos(B,D,F,H,I,L,N) using Dig-labeled cRNA probes specific for Gαolf (A,B), Omp (C,D), Nqo1 (E,F), Alk6 (G-J) and Msx1 (K-N). (A-F) Gαolf- and Omp-positive cells are present in all OE zones in control embryos(A,C) but show a restricted dorsomedial distribution in OE of Lhx2-/- embryos (arrows in B,D). The distribution of Gαolf- and Omp-positive cells in Lhx2-/- embryos is similar to the distribution of the dorsomedial-specific marker Nqo1 in control embryos (arrows in E). No Nqo1 expression over background is detected in Lhx2-/- embryos (F). (G-N) The gradient of Alk6expression (G,H; high dorsomedial and low ventrolateral) and the reverse gradient of Msx1 expression (K,L) are present in both control and Lhx2-/- embryos. High-power magnifications of Alk6 expression in the sustentacular cell layer (I,J) and Msx1 expression in the basal progenitor cell layer (M,N) is shown. Thus, cells expressing Alk6 and Msx1 show a similar distribution in OE cell layers in control and Lhx2-/-mice.

Fig. 5.

Abnormal region-specific OSN differentiation and normal region-specific gene expression in sustentacular and progenitor cells. In-situ hybridization analyses of serial coronal (dorsal up and medial left) sections of OE from control (A,C,E,G,I,K,M) and Lhx2-/- embryos(B,D,F,H,I,L,N) using Dig-labeled cRNA probes specific for Gαolf (A,B), Omp (C,D), Nqo1 (E,F), Alk6 (G-J) and Msx1 (K-N). (A-F) Gαolf- and Omp-positive cells are present in all OE zones in control embryos(A,C) but show a restricted dorsomedial distribution in OE of Lhx2-/- embryos (arrows in B,D). The distribution of Gαolf- and Omp-positive cells in Lhx2-/- embryos is similar to the distribution of the dorsomedial-specific marker Nqo1 in control embryos (arrows in E). No Nqo1 expression over background is detected in Lhx2-/- embryos (F). (G-N) The gradient of Alk6expression (G,H; high dorsomedial and low ventrolateral) and the reverse gradient of Msx1 expression (K,L) are present in both control and Lhx2-/- embryos. High-power magnifications of Alk6 expression in the sustentacular cell layer (I,J) and Msx1 expression in the basal progenitor cell layer (M,N) is shown. Thus, cells expressing Alk6 and Msx1 show a similar distribution in OE cell layers in control and Lhx2-/-mice.

Intact region-specific expression in sustentacular and progenitor cell layers

One possible scenario was that positional signals determining zone-specific gene expression in OE were absent or altered in Lhx2-/-embryos. The type I BMP receptor Alk6 is expressed in a dorsomedial to ventrolateral gradient by the sustentacular cells, whereas differential expression of the homeobox gene Msx1 defines a ventrolateral to dorsomedial counter gradient in the progenitor cell layer(Norlin et al., 2001). This spatially limited expression of Alk6 and Msx1 was unaltered in Lhx2-/- embryos(Fig. 5G-N). Since sustentacular and progenitor cells showed the expected regional differences in gene expression, the result strongly suggested that positional cues are present in OE of Lhx2-/- embryos.

In this study we show that the function of the LIM homeobox transcription factor Lhx2 is required for the generation of a heterogeneous population of individually and zonally specified OSNs. Lhx2 is apparently not required for the development of zonally specified progenitor cells or supporting cells of OE. Our results thus indicate that Lhx2 regulates a gene program that is intrinsic to OSN lineage that includes ORs and zone-specific genes. The neurons generated in the absence of Lhx2 function, and thereby also the functions of ORs and zone-specific genes, express pan-neuronal markers and have axons projecting toward the telencephalic vesicle. Progenitors thus apparently leave the cell cycle and acquire pan-neuronal traits characteristic of postmitotic neurons. We detect an accumulation of neurons that express the differentiation-promoting bHLH transcription factor Neurod1. Since OSNs normally co-express Neurod1 and pan-neuronal genes only transiently during differentiation (Cau et al.,1997), our results suggest that a majority of neurons in Lhx2-/- embryos correspond to immature neurons that have just begun to express pan-neuronal marker genes. Analyses of OR expression in the OE of adult animals indicate that there is a short delay period between the onset of pan-neuronal genes and the start of OR expression(Iwema and Schwob, 2003). Thus our results are compatible with the suggestion that neuronal differentiation in the Lhx2-/- embryos does not proceed past this delay period in an appropriate manner. In this regard it is interesting to note that accelerated differentiation as a consequence of OR expression has been suggested to underlie the feedback mechanism in which a given OR protein ensures high levels of expression of its corresponding gene(Lewcock and Reed, 2004). It is therefore conceivable that an Lhx2-dependent block in OR expression inhibits neuronal differentiation and/or prolongs the period during which OSN under normal circumstances are competent to select an OR gene for high-level expression. Even though the data does not allow us to determine if Lhx2 exerts its primary effect in progenitor cells or OSNs, it is interesting to note that Lhx2 has been identified by yeast one-hybrid screening as a protein that binds to conserved sequence motifs in the promoter regions of a defined subfamily of OR genes(Hoppe et al., 2003). Previous studies have demonstrated that promoter proximal regions of certain OR genes are sufficient to fully recapitulate proper expression in transgenic mice(Lewcock and Reed, 2004; Vassalli et al., 2002). Moreover, the feedback control mechanism of OR gene choice was shown to require a defined cis-acting locus control region with capacity to enhance transcription of OR genes belonging to different zonal sets(Lewcock and Reed, 2004). Similarly, we find that Lhx2 function appears required for expression of OR genes belonging to different zonal sets, which is compatible with the possibility that Lhx2 regulates transcription by binding to conserved DNA motifs in zonally expressed genes.

The OE in Lhx2-/- embryos shows an aberrant zonal distribution of neurons that express early (Neurod1 and pan-neuronal)and late (Gαolf and Omp) differentiation markers,respectively. This result suggests that neuronal differentiation and/or survival is under the influence of zonal factors and thus, that Lhx2is not required for all aspects of zonal specification of cells in OE. Evidence for the existence of zone-specific cues in the absence of Lhx2 is that the region-specific expression of Alk6 in sustentacular cells and Msx1 in progenitors is unperturbed in mutant embryos. Several other genes have been found to be expressed in progenitors in middle and ventrolateral zones exclusively(Tietjen et al., 2003). Interestingly, the expression of these genes in progenitors is either dependent or independent of Mash1 function. Since multipotent progenitors are present in OE it is tempting to speculate that a Mash1- and Lhx2-independent mechanism specify sustentacular cells to differentially express genes such as Alk6 in a zone-specific manner. Rbtn1 (Lmo1 – Mouse Genome Informatics), which belongs to the LIM only family, is among the transcription factors that were identified to be zonally expressed in a Mash1-dependent manner. LIM only factors interact with both LIM homeobox and bHLH proteins and can regulate Neurod1expression (Bao et al., 2000; Jurata et al., 1996). Thus,the zone-specific differentiation phenotype may be a consequence of lack of a zone-specific signaling protein in OSNs and/or absence of Lhx2function in progenitors. In either case it is likely that Lhx2deficiency accentuates a zone-specific difference in OSN maturation that is not as apparent in the OE of control animals. Evidence for regional differences under normal conditions has come from the finding that the regenerative capacity of OSNs in an adult animal differs between OE zones(Konzelmann et al., 1998). Moreover, the anterior OE, with a high proportion of the dorsomedial zone,appears to contain more OSNs that co-express early (Gap43) and late(Omp) differentiation markers compared with middle and ventrolateral zones (Iwema and Schwob,2003).

Collectively, the results presented show a role for Lhx2 in supplying a fundamental function in the regulated control of neuronal diversity within the OSN lineage. Further studies of Lhx2 function in the olfactory system will help clarify the precise role of Lhx2 in the gene regulatory mechanism, whereby an individual OSN selects a given response profile and acquires topographic precision in axon targeting.

During the revision of this manuscript, similar results were reported(Hirota and Mombaerts,2004).

We thank Dr Frank Margolis for the Omp antibody. This work was supported by grants from the Swedish Natural Science Research Council (B5101-1250/2001),the Swedish Cancer Society, the Wallenberg Foundation and the Vàsterbotten County.

Alenius, M. and Bohm, S. (
1997
). Identification of a novel neural cell adhesion molecule-related gene with a potential role in selective axonal projection.
J. Biol. Chem.
272
,
26083
-26086.
Alenius, M. and Bohm, S. (
2003
). Differential function of RNCAM isoforms in precise target selection of olfactory sensory neurons.
Development
130
,
917
-927.
Bao, J., Talmage, D. A., Role, L. W. and Gautier, J.(
2000
). Regulation of neurogenesis by interactions between HEN1 and neuronal LMO proteins.
Development
127
,
425
-435.
Berghard, A., Buck, L. B. and Liman, E. R.(
1996
). Evidence for distinct signaling mechanisms in two mammalian olfactory sense organs.
Proc. Natl. Acad. Sci. USA
93
,
2365
-2369.
Breitschopf, H., Suchanek, G., Gould, R. M., Colman, D. R. and Lassmann, H. (
1992
). In situ hybridization with digoxigenin-labeled probes: sensitive and reliable detection method applied to myelinating rat brain.
Acta Neuropathol.
84
,
581
-587.
Buck, L. and Axel, R. (
1991
). A novel multigene family may encode odorant receptors: a molecular basis for odor recognition.
Cell
65
,
175
-187.
Cau, E., Gradwohl, G., Fode, C. and Guillemot, F.(
1997
). Mash1 activates a cascade of bHLH regulators in olfactory neuron progenitors.
Development
124
,
1611
-1621.
Cau, E., Casarosa, S. and Guillemot, F. (
2002
). Mash1 and Ngn1 control distinct steps of determination and differentiation in the olfactory sensory neuron lineage.
Development
129
,
1871
-1880.
Guillemot, F., Lo, L. C., Johnson, J. E., Auerbach, A.,Anderson, D. J. and Joyner, A. L. (
1993
). Mammalian achaete-scute homolog 1 is required for the early development of olfactory and autonomic neurons.
Cell
75
,
463
-476.
Gussing, F. and Bohm, S. (
2004
). NQO1 activity in the main and the accessory olfactory systems correlates with the zonal topography of projection maps.
Eur. J. Neurosci.
19
,
1
-8.
Hirota, J. and Mombaerts, P. (
2004
). The LIM homeodomain protein Lhx2 is required for complete development of mouse olfactory sensory neurons.
Proc. Natl. Acad. Sci. USA
101
,
8751
-8755.
Holmfeldt, P., Brannstrom, K., Stenmark, S. and Gullberg, M.(
2003
). Deciphering the cellular functions of the Op18/Stathmin family of microtubule-regulators by plasma membrane-targeted localization.
Mol. Biol. Cell
14
,
3716
-3729.
Hoppe, R., Frank, H., Breer, H. and Strotmann, J.(
2003
). The clustered olfactory receptor gene family 262: genomic organization, promotor elements, and interacting transcription factors.
Genome Res.
13
,
2674
-2685.
Iwema, C. L. and Schwob, J. E. (
2003
). Odorant receptor expression as a function of neuronal maturity in the adult rodent olfactory system.
J. Comp. Neurol.
459
,
209
-222.
Iwema, C. L., Fang, H., Kurtz, D. B., Youngentob, S. L. and Schwob, J. E. (
2004
). Odorant receptor expression patterns are restored in lesion-recovered rat olfactory epithelium.
J. Neurosci.
24
,
356
-369.
Jurata, L. W., Kenny, D. A. and Gill, G. N.(
1996
). Nuclear LIM interactor, a rhombotin and LIM homeodomain interacting protein, is expressed early in neuronal development.
Proc. Natl. Acad. Sci. USA
93
,
11693
-11698.
Konzelmann, S., Saucier, D., Strotmann, J., Breer, H. and Astic,L. (
1998
). Decline and recovery of olfactory receptor expression following unilateral bulbectomy.
Cell Tissue Res.
294
,
421
-430.
Lewcock, J. W. and Reed, R. R. (
2004
). A feedback mechanism regulates monoallelic odorant receptor expression.
Proc. Natl. Acad. Sci. USA
101
,
1069
-1074.
Liem, K. F., Jr, Tremml, G. and Jessell, T. M.(
1997
). A role for the roof plate and its resident TGFbeta-related proteins in neuronal patterning in the dorsal spinal cord.
Cell
91
,
127
-138.
Miyawaki, A., Homma, H., Tamura, H., Matsui, M. and Mikoshiba,K. (
1996
). Zonal distribution of sulfotransferase for phenol in olfactory sustentacular cells.
EMBO J.
15
,
2050
-2055.
Monuki, E. S., Porter, F. D. and Walsh, C. A.(
2001
). Patterning of the dorsal telencephalon and cerebral cortex by a roof plate-Lhx2 pathway.
Neuron
32
,
591
-604.
Norlin, E. M., Alenius, M., Gussing, F., Hagglund, M., Vedin, V. and Bohm, S. (
2001
). Evidence for gradients of gene expression correlating with zonal topography of the olfactory sensory map.
Mol. Cell. Neurosci.
18
,
283
-295.
Porter, F. D., Drago, J., Xu, Y., Cheema, S. S., Wassif, C.,Huang, S. P., Lee, E., Grinberg, A., Massalas, J. S., Bodine, D. et al.(
1997
). Lhx2, a LIM homeobox gene, is required for eye,forebrain, and definitive erythrocyte development.
Development
124
,
2935
-2944.
Ressler, K. J., Sullivan, S. L. and Buck, L. B.(
1993
). A zonal organization of odorant receptor gene expression in the olfactory epithelium.
Cell
73
,
597
-609.
Ressler, K. J., Sullivan, S. L. and Buck, L. B.(
1994
). Information coding in the olfactory system: evidence for a stereotyped and highly organized epitope map in the olfactory bulb.
Cell
79
,
1245
-1256.
Sassoon, D. and Rosenthal, N. (
1993
). Detection of messenger RNA by in situ hybridization.
Methods Enzymol.
225
,
384
-404.
Schaeren-Wiemers, N. and Gerfin-Moser, A.(
1993
). A single protocol to detect transcripts of various types and expression levels in neural tissue and cultured cells: in situ hybridization using digoxigenin-labelled cRNA probes.
Histochemistry
100
,
431
-440.
Shirasaki, R. and Pfaff, S. L. (
2002
). Transcriptional codes and the control of neuronal identity.
Annu. Rev. Neurosci.
25
,
251
-281.
Sullivan, S. L., Bohm, S., Ressler, K. J., Horowitz, L. F. and Buck, L. B. (
1995
). Target-independet pattern specification in the olfactory epithelium.
Neuron
15
,
779
-789.
Tietjen, I., Rihel, J. M., Cao, Y., Koentges, G., Zakhary, L. and Dulac, C. (
2003
). Single-cell transcriptional analysis of neuronal progenitors.
Neuron
38
,
161
-175.
Vassalli, A., Rothman, A., Feinstein, P., Zapotocky, M. and Mombaerts, P. (
2002
). Minigenes impart odorant receptor-specific axon guidance in the olfactory bulb.
Neuron
35
,
681
-696.
Vassar, R., Ngai, J. and Axel, R. (
1993
). Spatial segregation of odorant receptor expression in the mammalian olfactory epithelium.
Cell
74
,
309
-318.
Vassar, R., Chao, S. K., Sitcheran, R., Nunez, J. M., Vosshall,L. B. and Axel, R. (
1994
). Topographic organization of sensory projections to the olfactory bulb.
Cell
79
,
981
-992.
Xu, Y., Baldassare, M., Fisher, P., Rathbun, G., Oltz, E. M.,Yancopoulos, G. D., Jessell, T. M. and Alt, F. W. (
1993
). LH-2: a LIM/homeodomain gene expressed in developing lymphocytes and neural cells.
Proc. Natl. Acad. Sci. USA
90
,
227
-231.
Yoshihara, Y., Kawasaki, M., Tamada, A., Fujita, H., Hayashi,H., Kagamiyama, H. and Mori, K. (
1997
). OCAM: A new member of the neural cell adhesion molecule family related to zone-to-zone projection of olfactory and vomeronasal axons.
J. Neurosci.
17
,
5830
-5842.
Youngentob, S. L., Kent, P. F. and Margolis, F. L.(
2003
). OMP gene deletion results in an alteration in odorant-induced mucosal activity patterns.
J. Neurophysiol.
90
,
3864
-3873.