Insm1 encodes a zinc-finger protein that is expressed in late progenitors and nascent neurons throughout the developing nervous system, including the embryonic OE [15, 16]. Here we show several lines of evidence indicating that, in embryonic OE, deletion of Insm1 causes a decrease in terminal, neuronogenic progenitors and an increase in earlier, proliferative progenitors. First, at certain embryonic stages the OEs of Insm1
mice (compared with OEs of Insm1
mice) contain more apical mitoses (E12.5) and fewer basal ones (E12.5 and E14.5). At these stages, progenitors undergoing mitosis apically generate more progenitors, whereas some progenitors undergoing mitosis at the basal lamina generate neurons. Second, at E12.5 the OEs of Insm1
mice display more cells expressing ASCL1 (a marker of TAPs, which migrate from apical to basal locations but still divide to produce progenitors) and fewer cells expressing NEUROD1 (a marker of INPs, which divide terminally next to the basal lamina to produce neurons). Third, by sequential incorporation of nucleoside analogs we demonstrate that, at E12.5, the OEs of Insm1
mice contain fewer progenitors that are terminally dividing. These results indicate that Insm1 promotes the transition of olfactory progenitors from apical, proliferative and uncommitted to basal, terminally dividing and neuron-producing.
Potential molecular mechanisms of Insm1 action
This study does not address the molecular mechanism by which Insm1 regulates the progression from early to late neuronal progenitors. In this regard, studies on EGL-46 offer no clues as its mechanism of action also has not been explored. Some insight, however, might be drawn from previous work revealing that Insm1 has at least two molecular mechanisms of action: it can bind promoter elements to repress transcription of certain genes and it can also promote cell cycle arrest and exit . The basic helix-loop-helix transcription factor ASCL1 binds to an element in the promoter of Insm1, and Ascl1 is required for Insm1 transcriptional activation in the sympathoadrenal lineage , ganglionic eminences  and OE . Furthermore, in the sympathoadrenal lineage, ganglionic eminences and cortex, ablation of Insm1 led to an up-regulation of Ascl1, and these results were interpreted as evidence that Insm1 represses Ascl1 [16, 23]. Similarly, the increase in ASCL1-expressing cells in the OE of Insm1
embryos here reported could result from direct transcriptional repression. However, this increase in ASCL1-expressing progenitors does not explain the decrease in basal progenitors and neurons, both of which require ASCL1.
Insm1 also binds an element in the NeuroD1 promoter and represses its transcription in pancreatic islet cell development [5, 46]. However, this regulatory interaction fails to explain the decrease of NEUROD1-expressing cells in the absence of Insm1. Alternatively, the retention of OE progenitors in a proliferative state in the absence of Insm1 might be more in keeping with the proposed role of this zinc-finger protein in cell cycle exit [5, 49].
Anatomical consequence of Insm1ablation: increase in apical and decrease in basal cells and neurons
Our results do explain the anatomical changes observed in Insm1-/- OEs. The increase in apical progenitors observed at E12.5 may account for the increase in cells observed in the apical layer at subsequent stages (E14.5 and E18.5). Many of these are likely sustentacular cells or their precursors because by E14.5 there is no longer an increase in apical progenitors. Likewise, the decrease of progenitors that are basal, NEUROD1-positive and terminally dividing at E12.5 may explain the decrease in nascent and mature neurons observed at this and later stages. Although apoptosis may also contribute to the decrease in neurons, its later onset (E14.5) and infrequent occurrence (100 times fewer cells appear undergoing programmed cell death than neurons are missing) indicates that it cannot be the sole or principal cause for neuronal decrease. The net increase in apical progenitors and cells and net decrease in basal progenitors and neurons seem to compensate for each other during the stages examined here because we never detected an overall difference in the size (measured in thickness, cell density or surface area in Figure 2 and Additional files 4 and 5) between Insm1
Potential roles of Insm1in cell survival and neuronal differentiation
The slight increase in cell death detected in Insm1
OEs might also be suggestive of a role of Insm1 in cell survival. This role would be modest, since it affects a very small fraction of Insm1
OE cells during the period examined (from E10.5 to E18.5). However, the increase in cell death is observed at a stage (E14.5) when olfactory receptor neurons innervate their targets in the olfactory bulb. In the wild type, the peak of cell death at this moment is thought to reflect the elimination of neurons that failed to establish proper synaptic contacts . Hence, given that the Insm1
embryos may also experience a reduction in olfactory bulb neurons (as they do in cortical neurons ), the slight increase in OE cell death may be an indirect consequence and not reflect a cell autonomous role of Insm1 in promoting neuronal survival.
The decrease in OMP-expressing, mature olfactory receptor neurons is more pronounced than the decrease in TuJ1-expressing, immature neurons (approximately 40% versus approximately 92%, respectively; Figure 3). The increase in cell death mentioned above could contribute to this difference if the cells dying are neurons prior to maturation. We were not able to elucidate which cells undergo programmed cell death in the OE, but previous studies have shown these to be TuJ1-expressing, young neurons . Although the increase of dying (that is, ACC3-expressing) cells (0.3 cells per 0.1 mm2 at E14.5) is ten-fold lower than the decrease in OMP-expressing neurons (3.1 cells per 0.1 mm2 at E14.5; after correcting for the 39% decrease in immature, TuJ1-expressing neurons), the transient nature of apoptotic cells underestimates their frequency. For example, if an apoptotic cell expresses ACC3 during a 3-hour period (a guess, since this has not been measured), then the density of missing cells (dying at a density of 0.3 cells per 0.1 mm2 mentioned above) would reach 3 per 0.1 mm2 in 30 hours, a value consistent with the observed reduction in OMP-expressing neurons.
The disproportionately larger reduction in OMP-expressing mature neurons than in TuJ1-expressing young neurons might also suggest a role for Insm1 in neuronal differentiation. In fact, Insm1 has been implicated in the differentiation of pancreatic and intestinal endocrine cells , adrenal chromaffin cells  and hindbrain monoaminergic neurons . However, the disproportionate reduction in OMP-expressing cells may simply reflect a more steep change in the onset of OMP expression than in the onset of TuJ1 expression during the observed stages (E14.5 and E18.5). Due to the lethality of Insm1
embryos, we cannot determine whether all TuJ1-expressing nascent neurons will mature to express OMP, or whether many of them will fail to differentiate (thus revealing a role for Insm1 in olfactory neuron differentiation).
Phenotypic similarities between Insm1 and egl-46mutants suggest functional conservation in neurogenic proliferation from nematodes to mammals
Mammalian Insm1 and nematode egl-46 have a similar pattern of expression in neuronal lineages: both are transiently expressed in neuronogenic progenitors and nascent neurons and not in many of the earlier, proliferative progenitors (Figure 10A, C). In particular, nematode egl-46 is expressed in progenitors of the motorneuron and Q-neuroblast lineages undergoing N/N but not P/P or P/N divisions, whereas mouse Insm1 is expressed in OE progenitors undergoing mitosis basally (many of which are neuronogenic) but not apically (many of which are proliferative). Although our data cannot determine whether Insm1 is exclusively expressed in terminally dividing N/N progenitors of OE, they indicate that Insm1 is expressed in late neuronogenic progenitors and not in many of the earlier, proliferative progenitors. In nematodes, mutation of egl-46 results in a transient decrease in neuronogenic (N/N) divisions and a concomitant increase in proliferative (P/N or P/P) divisions (Figure 10B). This occurs because progenitors that by lineage are expected to terminally divide fail to do so, and instead some of their progeny divide one or more times. Similarly we find that, in the mammalian OE, deletion of Insm1 results in fewer terminally dividing neuronogenic progenitors and in a concomitant increase in proliferative (apical and ASCL1-expressing) progenitors (Figure 10D). Hence, we propose that Insm1 and egl-46 play evolutionarily conserved roles in promoting terminal, neuronogenic divisions.
It should be noted that, in both organisms, the mutant effects are modest. In egl-46 mutants, only some terminal progenitors are delayed, and only by one or two cell divisions [2–4]. In Insm1 mutants, the increase in apical progenitors is transient (detected at E12.5 but not at E10.5 or E14.5) and small (42% increase; Figure 6), as is the decrease in basal progenitors (45% at E12.5 and 35% at E14.5; Figure 6). The subtlety of both phenotypes demonstrates that these two genes are not essential for the transition from proliferative to neuronogenic progenitors, and instead suggests that their function is to fine tune their occurrence.
A difference between the egl-46(loss-of-function) and Insm1
phenotypes is that, in mutant nematodes, an increase in proliferative divisions eventually results in additional neurons, whereas mouse Insm1
OEs still contain fewer neurons than the wild type at the latest stages examined (E18.5). The embryonic lethality of Insm1
prevents determining whether more neurons would be eventually produced. However, in these mutant OEs we also detected an increase in apical cells, which would eventually become sustentacular glia. Because apical progenitors stop transitioning basally by E16.5 [1, 25], the increase in apical progenitors observed at E12.5 may eventually result in more sustentacular cells and not in more neurons.
Phenotypic and other similarities suggest a common mechanism of neurogenic proliferation between OE and cortex
In embryonic cortex, progenitors undergoing mitosis at the apical side of the ventricular zone generate more progenitors, some of which migrate in the basal direction. These basally located progenitors divide one or more times to produce neurons, which migrate to the cortical plate and differentiate [50–55]. The effects of Insm1 deletion in OE reported here share similarities with those reported in dorsal telencephalon . In both embryonic OE and cortex, Insm1 deletion results in an enlargement of the area with apical progenitors (the ventricular zone in cortex and the apical layer in OE), a decrease in basal, neuronogenic progenitors (the intermediate progenitor cells of the cortical subventricular zone and the INPs of OE), and a decrease in neurons. Furthermore, both neuroepithelia share striking similarities that extend beyond the Insm1
phenotype. At embryonic stages, both have progenitors dividing apically (near the airways in OE and near the ventricle in cortex), some of which display radial glia characteristics . These apical progenitors undergo interkinetic nuclear migration and generate glial-like epithelial cells (ependymal in cortex and sustentacular in OE) as well as additional progenitors that migrate basally (to the basal lamina in OE and to the subventricular zone in cortex). In both embryonic neuroepithelia, many of these basal progenitors divide terminally to produce neurons. After embryogenesis, proliferation subsides in apical areas but is maintained in basal areas (subventricular zone and base of OE), where adult neural stem cells reside and neurogenesis occurs. All these similarities suggest that both neuroepithelia share a common mechanism of progenitor migration, proliferation and differentiation orchestrating neurogenesis and gliogenesis. We propose that common molecular pathways involving Insm1 and other factors may direct the development of OE and cortex and, perhaps, other neural tube and placodal neuroepithelia.