|Home | About | Journals | Submit | Contact Us | Français|
Neuronal migration is essential for the development of the mammalian brain. Here, we document severe defects in neuronal migration and reduced numbers of neurons in lamin B1–deficient mice. Lamin B1 deficiency resulted in striking abnormalities in the nuclear shape of cortical neurons; many neurons contained a solitary nuclear bleb and exhibited an asymmetric distribution of lamin B2. In contrast, lamin B2 deficiency led to increased numbers of neurons with elongated nuclei. We used conditional alleles for Lmnb1 and Lmnb2 to create forebrain-specific knockout mice. The forebrain-specific Lmnb1- and Lmnb2-knockout models had a small forebrain with disorganized layering of neurons and nuclear shape abnormalities, similar to abnormalities identified in the conventional knockout mice. A more severe phenotype, complete atrophy of the cortex, was observed in forebrain-specific Lmnb1/Lmnb2 double-knockout mice. This study demonstrates that both lamin B1 and lamin B2 are essential for brain development, with lamin B1 being required for the integrity of the nuclear lamina, and lamin B2 being important for resistance to nuclear elongation in neurons.
The patterning of the cerebral cortex during embryonic development involves the rapid expansion of the pool of neuronal progenitors in the ventricular zone and their radial migration as neurons into the cortical plate (Gupta et al., 2002 ; Ayala et al., 2007 ; Wynshaw-Boris, 2007 ). Successive waves of neurons migrate to form distinct layers within the cortical plate, with each new layer located more superficially than earlier layers (Gupta et al., 2002 ). The positioning of neurons into cortical layers is crucial for their identity and their ability to form proper connections within the brain (Wynshaw-Boris, 2007 ), and defects in neuronal migration have catastrophic consequences for the organization of the brain (Gupta et al., 2002 ; Ayala et al., 2007 ; Wynshaw-Boris, 2007 ). The study of neurodevelopmental defects in humans has helped to identify cytoplasmic proteins required for neuronal migration (e.g., LIS1, NDE1, NDEL1; Gupta et al., 2002 ; Ayala et al., 2007 ; Wynshaw-Boris, 2007 ). These proteins regulate a network of microtubules and dynein motors and are essential for moving the cell nucleus in the direction of the leading edge of the neuron—a process called nuclear translocation (Vallee and Tsai, 2006 ; Tsai et al., 2007 ; Wynshaw-Boris, 2007 ). Defects that prevent nuclear translocation block the migration of neurons to the cortical plate and lead to severe neurodevelopmental abnormalities (Solecki et al., 2006 ; Wynshaw-Boris, 2007 ).
Nuclear translocation depends on effective connections between the cell nucleus and the microtubule network in the cytoplasm (Tanaka et al., 2004 ; Wynshaw-Boris, 2007 ; Pawlisz et al., 2008 ). Although the cytoplasmic factors regulating neuronal migration have been thoroughly investigated, it is only recently that nuclear proteins with roles in this process have been identified. First, Zhang et al. (2009) identified key roles for the SUN proteins and nesprins in neuronal migration. These proteins of the nuclear envelope are components of the LINC complex, a molecular bridge that links the nucleus to the cytoskeleton (Fridkin et al., 2009 ; Méjat and Misteli, 2010 ). Zhang et al. (2009) documented abnormalities in neuronal migration in mice lacking both SUN1 and SUN2 and in mutant mice expressing inactive forms of Syne-1/Nesprin 1 and Syne-2/Nesprin 2. Shortly thereafter, Coffinier et al. (2010a) discovered that lamin B2, a protein of the nuclear lamina, is critical for neuronal migration during embryonic development. In Lmnb2-knockout mice (Lmnb2–/–), the layering of neurons within the cerebral cortex was abnormal, and the cerebellum was reduced in size and devoid of folds (Coffinier et al., 2010a ). The latter observations provided the first indication that the nuclear lamina is important for the development of the mammalian brain.
In addition to lamin B2, the nuclear lamina of most differentiated somatic cells contains lamins A, C, and B1 (Dechat et al., 2008 ). There is little reason to believe that LMNA, the gene encoding lamins A and C, is crucial for the development of the brain. Lmna is expressed late in mouse development (Rober et al., 1989 ), and Lmna-deficient mice develop normally, surviving for 5–6 wk before succumbing to cardiomyopathy and/or muscular dystrophy (Sullivan et al., 1999 ). Hundreds of clinically significant mutations have been identified in LMNA, but none have been linked to neurodevelopmental abnormalities. Most LMNA mutations in humans cause muscular dystrophy and/or cardiomyopathy, but some cause lipodystrophy, peripheral neuropathy, and progeria (Worman et al., 2009 ).
Whether lamin B1 is involved in brain development has been an open question. Vergnes et al. (2004) reported that lamin B1–deficient (Lmnb1Δ/Δ) mice were small, exhibited abnormalities in the lungs and skeleton, and died soon after birth. Lmnb1Δ/Δ fibroblasts manifest nuclear blebbing and early senescence (Vergnes et al., 2004 ). The possibility of CNS disease in Lmnb1Δ/Δ embryos was not investigated, but the mutant mice had an abnormally shaped cranium (Vergnes et al., 2004 ). This finding, together with the neurodevelopmental abnormalities in Lmnb2–/– mice, suggested that lamin B1 might be important for brain development. In the present study, we investigated that possibility, taking advantage of both the original Lmnb1Δ/Δ mouse model and newly developed conditional knockout alleles for Lmnb1 and Lmnb2 (Yang et al., 2011 ).
The abnormally shaped cranium in Lmnb1Δ/Δ embryos (Vergnes et al., 2004 ) led us to consider a potential role for lamin B1 in brain development. Histological analyses revealed that the brains of newborn Lmnb1Δ/Δ mice were abnormal (Figure 1). The layering of neurons in the cerebral cortex was absent, with reduced numbers of cells (Figure 1, A–C); no lamination was observed in the hippocampus; and the cerebellum was reduced in size, with no foliation (Figure 1A). The number of cortical neurons was reduced, as judged by immunostaining for the neuronal marker NeuN (Figure 1C). The neurodevelopmental abnormalities in Lmnb1Δ/Δ mice suggested that Lmnb1 is important during embryonic development. Indeed, β-galactosidase staining at E15.5 revealed Lmnb1 expression throughout the cerebral cortex (Figure 1D). At the same stage, Lmnb2 expression was prominent in the ventricular zone of the cortex (Coffinier et al., 2010a ), and Lmna expression in the brain was minimal (although it was detected in the surrounding mesenchyme; Figure 1D).
At E15.5 and E17.5, the cortical plate was thinner in Lmnb1Δ/Δ embryos than in wild-type (WT) embryos (Figure 2, A–C). Immunostaining for TBR1 (a marker of cortical layer VI) revealed similar numbers of TBR1-positive (TBR1+) cells in WT and Lmnb1Δ/Δ embryos at E13.5; however, there were fewer TBR1+ cells in Lmnb1Δ/Δ embryos by E15.5, and those neurons were located more superficially than in WT brains (Figure 2D). Immunohistochemical studies of E16.5 embryos with antibodies against Otx1 (a marker of cortical layers V–VI) and TBR1 revealed that neurons expressing those markers were located more superficially in Lmnb1Δ/Δ brains (Figure 2E). The positions of the subplate and of the lateral projections, visualized by staining with the monoclonal antibodies CS56 and L1, respectively, were also more superficial than normal (Figure 2E). At E18.5, neurons expressing the layer V marker Ctip2 were also located more superficially in Lmnb1Δ/Δ embryos (Figure 2F). The abnormal positioning of those deep-layer neurons suggested a defect of the neurons born later in forming the upper layers of the cortex. To test this hypothesis, we performed neuronal birthdating experiments; pregnant mice were injected with bromodeoxyuridine (BrdU) at E13.5, and BrdU-labeled neurons were examined in E18.5 embryos (Figure 2F). In WT brains, BrdU-positive neurons (i.e., cells born at E13.5) were found in layer V, whereas in the brain of Lmnb1Δ/Δ mice, BrdU-positive neurons were scattered throughout the cortical plate (Figure 2F). Together, these studies demonstrated a defect in neuronal migration in Lmnb1Δ/Δ embryos. We used immunohistochemistry to assess Reelin expression in the marginal zone, as Reelin deficiency is known to impair neuronal migration (Rice and Curran, 2001 ). However, Reelin appeared to be expressed similarly in Lmnb1Δ/Δ and control embryos at E12.5–E17.5 (Supplemental Figure S1).
The small size of the cortical plate was due in part to reduced numbers of neuronal progenitors. At E13.5, similar numbers of Sox2+ progenitors were found in Lmnb1Δ/Δ and WT brains, but by E14.5–E15.5 their numbers were clearly reduced in Lmnb1Δ/Δ brains (167 ± 10 Sox2+ cells in E15.5 mutant embryos [n = 3] vs. 267 ± 72 in WT embryos [n = 3], per area of 430 × 470 μm; p = 0.06; Figure 3A). At the same stage, the proportion of Sox2+ cells expressing the mitotic marker Ki67 was higher in Lmnb1Δ/Δ brains, suggesting the possibility that neurons in Lmnb1Δ/Δ embryos spend more time in the S–M phase (Figure 3A). At E16.5, the numbers of Ki67+Sox2+ cells were ~20% higher in Lmnb1Δ/Δ embryos than in WT embryos (47.8 ± 5.2% [n = 3] vs. 36.7 ± 4.6% [n = 4]; p = 0.03). In addition to producing cortical neurons, neuronal progenitors give rise to intermediate progenitors that accumulate in the subventricular zone and express TBR2 (Dehay and Kennedy, 2007 ). Intermediate progenitors differentiate at later stages and contribute to layers II–III of the cortical plate. At E15.5 and E17.5, we observed fewer TBR2+ cells in the subventricular zone of Lmnb1Δ/Δ embryos (Figure 3B). At E15.5, the number of intermediate progenitors was reduced by 50% in Lmnb1Δ/Δ embryos (163 ± 52 TBR2+ cells in an area of 430 × 350 μm; at least three areas evaluated per embryo; n = 3 embryos) compared with WT embryos (322 ± 49 cells; n = 3 embryos; p = 0.018). Aside from reduced proliferation, we detected apoptotic cells in the cortex of E16.5 Lmnb1Δ/Δ embryos by staining for active caspase 3 (Figure 3C). The brains of E16.5 Lmnb2–/– embryos also stained positively for active caspase 3, but the apoptotic cells were fewer in number and confined to the cortical plate; in contrast, fewer than two positive cells were observed per slice of WT cortex (Figure 3C).
Defects in lamins A and C often lead to severe nuclear shape abnormalities in cultured fibroblasts (Muchir et al., 2004 ), but misshapen nuclei are seldom found in mouse tissues. For example, we observed many nuclear blebs in fibroblasts cultured from “lamin A–only mice” (LmnaLAO/LAO), but no misshapen nuclei were found in the tissues of those mice (Coffinier et al., 2010b ). Vergnes et al. (2004) documented nuclear blebs in Lmnb1Δ/Δ fibroblasts, but we were skeptical that we would find misshapen nuclei in tissues of Lmnb1Δ/Δ mice. To our surprise, however, we observed severe nuclear shape abnormalities in the cerebral cortex of E16.5 Lmnb1Δ/Δ embryos. In brain sections stained for the nuclear envelope protein Lap2β or for lamin B2, 24.8 ± 6.8% of cortical neurons from Lmnb1Δ/Δ embryos (n > 350 cells evaluated per embryo, three different embryos) contained a solitary nuclear bleb versus none in WT neurons (n > 123 cells evaluated per embryo, three different embryos; p = 0.003; Figure 4A). Immunostaining for lamin B2 uncovered a second abnormality: 75 ± 6.1% of the cortical neurons in Lmnb1Δ/Δ embryos exhibited an asymmetric distribution of lamin B2 at the nuclear rim (Figure 4A; compared with none in the WT samples; same numbers of cells evaluated; n = 3 embryos/group; p < 0.0001). The nuclear bleb was invariably found in the region of the nuclear rim enriched in lamin B2 (Figure 4A).
Lmnb2–/– fibroblasts do not have nuclear blebs (Coffinier et al., 2010a ), but the finding of misshapen nuclei in Lmnb1Δ/Δ neurons prompted us to investigate whether Lmnb2–/– neurons might also have nuclear shape abnormalities (Figure 4B). At E16.5, cortical neurons of Lmnb2–/– mice did not have nuclear blebs, but we found cells with an elongated nucleus in lamin B2–deficient brains (Figure 4B), and there was a significant increase in the length of the nucleus in lamin B2–deficient embryonic neurons in situ compared with WT neurons (p < 0.0001; Supplemental Figure S2A). Lamin B1 was evenly distributed at the nuclear rim of Lmnb2–/– neurons—even in cells with elongated nuclei (Figure 4B).
Nuclear shape abnormalities were also observed in neurons grown from cortical explants of E13.5 embryos. Many Lmnb1Δ/Δ neurons had a single nuclear bleb, and some exhibited an asymmetric distribution of lamin B2 (Supplemental Figure S3). In the case of neurons from Lmnb2–/– embryos, we observed occasional comet-shaped nuclei with detached centrosomes (located >20 μm from the cell nucleus; Supplemental Figure S2B). Comet-shaped nuclei were never observed in neuronal progenitors from WT embryos.
Lmnb1Δ/Δ and Lmnb2–/– mice die soon after birth. To assess postnatal phenotypes in the brain, we used Lmnb1 and Lmnb2 conditional knockout alleles (Yang et al., 2011 ) and the Emx1-Cre transgene (Gorski et al., 2002 ) to generate forebrain-specific Lmnb1- and Lmnb2-knockout mice (Emx1-Cre Lmnb1fl/fl and Emx1-Cre Lmnb2fl/fl, respectively). For both conditional knockout alleles, CRE-mediated recombination excises exon 2, creating a frameshift and yielding a null allele. Specific expression of the Emx1-Cre transgene in the forebrain was documented with a CRE-activated lacZ reporter (Supplemental Figure S4A; Soriano, 1999 ), confirming results already in the literature (Gorski et al., 2002 ). Efficient forebrain-specific gene inactivation during embryogenesis was achieved with both the Lmnb1 and Lmnb2 conditional alleles. Immunohistochemical studies on the brain from an E15.5 Emx1-Cre Lmnb1fl/fl embryo revealed markedly reduced lamin B1 expression in the forebrain (Supplemental Figure S4B). Higher-magnification images of Emx1-Cre Lmnb1fl/fl and Emx1-Cre Lmnb2fl/fl brains revealed that ~90% of E15.5 cortical cells had no lamin B1 or lamin B2, respectively (Supplemental Figure S4C).
Emx1-Cre Lmnb1fl/fl and Emx1-Cre Lmnb2fl/fl mice were born at the expected Mendelian frequency, appeared to have normal longevity (mice were observed for >1 yr), and were grossly indistinguishable from WT mice. After removal of the skin, however, we observed that the cranium in both models was smaller, and the cerebral cortex was reduced in size (Figure 5, A and B). At 4 mo of age, the average length of the cortex in Emx1-Cre Lmnb1fl/fl mice was 0.65 ± 0.03 cm versus 0.90 ± 0.03 cm in WT mice (n = 3 per group; p < 0.001). The length of the cortex in Emx1-Cre Lmnb2fl/fl mice was 0.80 ± 0.0 cm versus 1.02 ± 0.02 cm in WT mice (n = 2 per group). We also bred mice lacking both lamin B1 and lamin B2 in the forebrain (Emx1-Cre Lmnb1fl/fl Lmnb2fl/fl). The cortex of these “double-knockout” mice was significantly smaller than those of Emx1-Cre Lmnb1fl/fl and Emx1-Cre Lmnb2fl/fl mice (Figure 5, C and D). Compared to WT siblings, the double-knockout mice had very similar body weights (16.20 ± 2.96 g vs. 16.10 ± 2.39 g) at 1 mo of age, but the brain weight of the double-knockout mice was significantly smaller than that of WT mice (0.27 ± 0.01 vs. 0.47 ± 0.02 g, respectively; n = 4 and 5 females; p < 0.0001). Brain sections of 1-mo-old double-knockout mice revealed atrophy of the cortex, which was reduced to a thin layer of tissue overlaying the striatum (Figure 5, E and F). Coronal sections also showed a complete absence of the hippocampal structures (Figure 5F). Immunohistochemistry studies on E17.5 Emx1-Cre Lmnb1fl/fl Lmnb2fl/fl embryos revealed cells that lacked both lamin B1 and lamin B2 (Supplemental Figure S4D). Immunohistochemical studies on the thin layer of tissue above the striatum in adult Emx1-Cre Lmnb1fl/fl Lmnb2fl/fl mice failed to detect any neurons (Supplemental Figure S4E), and none of the remaining cells lacked expression of both lamin B1 and lamin B2 (Supplemental Figure S4F).
We analyzed the effect of the forebrain-specific inactivation of Lmnb1 or Lmnb2 on the layering of neurons in the cortex. As expected, the layering of the cortical neurons in the adult brain was abnormal in both Emx1-Cre Lmnb1fl/fl mice and Emx1-Cre Lmnb2fl/fl mice. Immunostaining for the neuronal marker NeuN revealed reduced numbers of neurons in Emx1-Cre Lmnb1fl/fl mice, with most neurons expressing the layer V marker Ctip2 and very few neurons expressing the layer II–III marker Cux1 (Figure 5G). Emx1-Cre Lmnb2fl/fl mice had fewer cortical neurons than did WT mice (Figure 5G and Supplemental Figure S4C), and neurons failed to organize into proper layers. However, there were significantly more neurons and more Cux1-positive cells in Emx1-Cre Lmnb2fl/fl brains than in Emx1-Cre Lmnb1fl/fl brains (Figure 5G).
Misshapen cell nuclei were easily detectable in both forebrain-specific knockout models. Many neurons of Emx1-Cre Lmnb1fl/fl embryos contained a solitary nuclear bleb and exhibited an asymmetric distribution of lamin B2 (Supplemental Figure S5A). In Emx1-Cre Lmnb2fl/fl brains, we observed an increased frequency of neurons with elongated nuclei, but lamin B1 was distributed evenly at the nuclear rim (Supplemental Figure S5B). Of interest, few neurons in the cerebral cortex of adult Emx1-Cre Lmnb1fl/fl and Emx1-Cre Lmnb2fl/fl mice exhibited abnormal nuclear morphology (Figure 6A). However, in the dentate gyrus of adult Emx1-Cre Lmnb1fl/fl mice, many neurons exhibited a markedly asymmetric distribution of lamin B2 (Figure 6B). In contrast, neurons in the dentate gyrus from Emx1-Cre Lmnb2fl/fl mice exhibited normal nuclear morphology (Figure 6C).
Recently Yang et al. (2011) showed that the absence of both lamin B1 and lamin B2 in skin keratinocytes had no effect on skin development or on keratinocyte proliferation and survival. In contrast, we found dramatic nuclear abnormalities in neurons and severe neurodevelopmental abnormalities in Lmnb1- and Lmnb2-deficient mice. In addition, no neurons were detected in the remnant of cortex found in adult Emx1-Cre Lmnb1fl/fl Lmnb2fl/fl mice. We suspected that the different effects of lamin B deficiencies in skin and brain might relate to different levels of lamin A/C expression. Indeed, Lmna expression is almost undetectable in the embryonic brain, but Lmna expression in the skin of E15.5 and E19.5 embryos is robust (Supplemental Figure S6A). Immunostaining of embryos at E17.5 provided further evidence for robust lamin A/C expression in the skin; however, no lamin A/C expression could be detected in the cortical plate (Supplemental Figure S6B).
In contrast to the embryonic brain, which does not express significant amounts of lamin A/C (Figure 1D), all three lamin genes are expressed in the adult mouse brain (Supplemental Figure S6C). In the setting of lamin B1 deficiency, a robust signal for lamin A/C was associated with a normal distribution of lamin B2 at the nuclear rim (Supplemental Figure S6D). Of interest, some neurons of the dentate gyrus in adult mice—which retained an abnormal distribution of lamin B2 (Figure 6C and Supplemental Figure S6E)—exhibited lower levels of lamin A/C expression (Supplemental Figure S6E).
The phenotypes of the conventional knockout mice and the forebrain-specific knockout mice are summarized in Table 1.
In the present study, we found that proper development of the brain depends on lamin B1. Lmnb1Δ/Δ embryos have reduced numbers of neurons and abnormal layering of neurons in the cerebral cortex, as well as abnormalities in the hippocampus and cerebellum. The disorganization of cortical neurons in Lmnb1Δ/Δ embryos was caused by impaired neuronal migration, as judged by birthdating and immunohistochemical studies. In wild-type embryos, the neurons born at E13.5 (i.e., BrdU-labeled neurons) were detected 5 d later within layer V of the cortical plate, but in Lmnb1Δ/Δ embryos, many BrdU-labeled neurons were found deeper in the cortex, indicating defective migration. In addition, in Lmnb1Δ/Δ embryos, immunohistochemical markers for cortical layers V and VI were found in more superficial positions in the cortex. The neurodevelopmental abnormalities in Lmnb1Δ/Δ embryos were more severe than those described in Lmnb2–/– embryos (Coffinier et al., 2010a ). Consistent with those findings, the phenotypes of forebrain-specific Lmnb1-knockout mice (Emx1-Cre Lmnb1fl/fl) were more severe than those of forebrain-specific Lmnb2-knockout mice (Emx1-Cre Lmnb2fl/fl). The presence of neurodevelopmental abnormalities in Emx1-Cre Lmnb1fl/fl mice should allay any previous concerns that the developmental defects in Lmnb1Δ/Δ brains resulted from toxic effects of the lamin B1–β-galactosidase fusion protein produced by the Lmnb1Δ allele (Vergnes et al., 2004 ). Emx1-Cre Lmnb1fl/fl mice have a true null allele, yet they have the same neuropathology found in Lmnb1Δ/Δ mice (e.g., disorganization of cortical layers, paucity of neurons, reduced size of the forebrain).
The reduced size of the cortex in Lmnb1Δ/Δ mice was due to reduced numbers of neuronal progenitors. This defect is likely due in part to defective cell proliferation, but this defect may not be unique to neuronal progenitors, as the entire Lmnb1Δ/Δ embryo—and not just the brain—is small (Vergnes et al., 2004 ). In addition to reduced cell proliferation, apoptosis contributes to the reduced cell density, as we detected widespread activation of caspase 3 in the forebrain of Lmnb1Δ/Δ embryos. Apoptosis was previously reported in association with other defects that impair neuronal migration—for example loss-of-function mutations in LIS1 (Yingling et al., 2008 ), NDEL1 (Sasaki et al., 2005 ), NDE1 (Feng and Walsh, 2004 ), and cyclin-dependent kinase 5 (Li et al., 2002 ). However, we cannot exclude a more general effect of lamin B1 deficiency on cell viability. Earlier studies with cultured cells suggested that B-type lamins could play important roles in DNA replication, mitotic spindle assembly, heterochromatin organization, and regulation of gene expression (Moir et al., 2000 ; Lopez-Soler et al., 2001 ; Hutchison, 2002 ; Tsai et al., 2006 ; Dechat et al., 2009 ). These functions might be crucial in the developing brain, where the expression of A-type lamins is virtually absent (Rober et al., 1989 ).
In an earlier study (Coffinier et al. 2010a ), we found that the size of the cerebral cortex was only slightly smaller in E16.5–19.5 Lmnb2–/– embryos than in wild-type embryos. However, the forebrain size of adult Emx1-Cre Lmnb2fl/fl mice was much smaller than the forebrain of wild-type mice, presumably because brain development proceeds for weeks after birth, and the full effects of lamin B2 deficiency on cell proliferation and survival are therefore more manifest in older mice. The impact of B-type lamins on cell viability was even more dramatic in forebrain-specific Lmnb1/Lmnb2–knockout mice. During embryonic development, it was possible to identify neurons lacking both lamin B1 and lamin B2 in the cortical plate of Emx1-Cre Lmnb1fl/fl Lmnb2fl/fl mice. By 1 mo of age, however, the forebrain of Emx1-Cre Lmnb1fl/fl Lmnb2fl/fl mice was severely atrophic, and it was impossible to find any Lmnb1/Lmnb2–deficient neurons.
The abnormal nuclear morphology in lamin B1– and lamin B2–deficient neurons points to the reduced integrity of the nuclear envelope as a likely explanation for defective neuronal migration. The nuclei of many neurons in Lmnb1Δ/Δ embryos contained a nuclear bleb and exhibited an asymmetric distribution of lamin B2, whereas Lmnb2–/– neurons had an increased frequency of elongated nuclei. Earlier studies identified misshapen nuclei in Lmnb1Δ/Δ fibroblasts (Vergnes et al., 2004 ) and in HeLa cells, where Lmnb1 expression had been knocked down with small interfering RNAs (Shimi et al., 2008 ). In Lmnb1Δ/Δ fibroblasts, the nucleus often contains multiple blebs, whereas Lmnb1Δ/Δ neurons contain a single bleb. We suggest that the “solitary-bleb” phenotype might result from unidirectional forces applied on the nucleus during nuclear translocation. In contrast, the multiple blebs in fibroblasts might reflect multidirectional strain from the cytoskeleton as the cells spread out on the culture dish. We did not find nuclear blebs in Lmnb2–/– fibroblasts (Coffinier et al., 2010a ) or in the neurons of lamin B2–deficient mice, but we did observe elongated, “comet-shaped” nuclei in the cortical neurons of Lmnb2–/– and Emx1-Cre Lmnb2fl/fl embryos. The stretched-out nuclei in cultured Lmnb2–/– neurons were often associated with a markedly displaced centrosome, suggesting that nuclear abnormalities in lamin B2–deficient neurons were linked to defective nuclear translocation. Lamin B1 was distributed evenly at the nuclear rim in lamin B2–deficient neurons.
In the present study, we demonstrated that lamin B1 deficiency is sufficient to cause severe neurodevelopmental abnormalities. In contrast, Yang et al. (2011) created mice lacking both lamin B1 and lamin B2 in skin keratinocytes and found that skin histology and keratinocyte proliferation were entirely normal. Thus, the consequences of losing the expression of the B-type lamins are significantly different in the brain and skin. Our experiments uncovered a likely mechanism for these differences—strikingly different levels of lamin A/C expression. In the skin of mouse embryos, Lmna expression is robust, as judged by β-galactosidase staining or immunohistochemistry, whereas it is minimal in the developing brain. Strong expression of Lmna in Lmnb1/Lmnb2–deficient skin is associated with protection from both skin pathology and nuclear shape abnormalities, as judged by immunohistochemical studies on frozen sections of skin. In contrast, the absence of Lmna expression in neurons of lamin B1–deficient embryos is associated with severe neuropathology and striking nuclear shape abnormalities and an asymmetric distribution of lamin B2 at the nuclear rim.
Abnormalities in nuclear shape and lamin B2 distribution were apparent in the cerebral cortex neurons of both Lmnb1Δ/Δ and Emx1-Cre Lmnb1fl/fl embryos. However, the nuclear abnormalities were absent in cortical neurons of 1-mo-old Emx1-Cre Lmnb1fl/fl mice, when Lmna expression is robust. Of interest, the asymmetric distribution of lamin B2 persisted in the dentate gyrus of adult Emx1-Cre Lmnb1fl/fl mice, where Lmna expression remains extremely low. Thus, it appears that proper lamin B2 distribution in neurons requires the expression of either lamin B1 or lamin A/C.
Because the nuclear lamina is located within the nucleus, a molecular bridge between the cytoplasmic network of microtubules and the nuclear lamina is almost certainly required to mediate nuclear translocation. Zhang et al. (2009) showed that the SUN proteins and nesprins are required for neuronal migration, and they proposed that the SUN proteins could interact with “lamin B” in the nuclear lamina. This hypothesis seems reasonable, particularly in light of recent studies showing interactions between SUN proteins and A-type lamins in nuclear movement in fibroblasts (Folker et al., 2011 ). Studies with Lmnb1Δ/Δ fibroblasts lend additional support for a role of B-type lamins in linking the nucleus to the cytoskeleton. Videomicroscopy studies of Lmnb1Δ/Δ fibroblasts revealed that the nucleus spins within the cell (Ji et al., 2007 ), suggesting that lamin B1 is crucial for tethering the nucleus to the cytoskeleton.
Finding neurodevelopmental abnormalities in lamin B1– and lamin B2–deficient mice provides new insights into the function of the nuclear lamina. Until recently, most investigators focused on A-type lamins and their involvement in muscular dystrophy, cardiomyopathy, lipodystrophy, and progeria (Worman et al., 2009 ). The prevailing view was that B-type lamins are essential proteins in eukaryotic cells, crucial for mitosis and many other vital functions in the cell nucleus (Dechat et al., 2008 ; Worman et al., 2009 ), but our present studies point to specific roles for the B-type lamins in brain development. It remains to be determined whether lamins B1 and B2 have unique or overlapping functions in neuronal migration and brain development. In addition, the partners for lamin B1 and lamin B2 need to be established, along with the structural features of these lamins that are crucial for their function in the brain. In particular, we wonder whether the conserved farnesyl–lipid anchors on lamin B1 and lamin B2 might play a role in keeping the nuclear lamina attached to the inner nuclear membrane during neuronal migration. Finally, our work heralds a new path for investigating the human genetics of nuclear lamins. Neuronal migration defects underlie not only lethal neurodevelopmental defects such as lissencephaly (Wynshaw-Boris and Gambello, 2001 ; Wynshaw-Boris, 2007 ), but also other neurological diseases, such as epilepsy, schizophrenia, and autism (Kahler et al., 2008 ; Deutsch et al., 2010 ; Wegiel et al., 2010 ). We suspect that whole-exome sequencing on patients with severe neurological diseases will eventually uncover nonsense and missense mutations in LMNB1 and LMNB2.
The Lmnb1Δ allele (Vergnes et al., 2004 ) contains a βgeo gene-trap cassette in intron 5 of Lmnb1 and produces a lamin B1–βgeo fusion protein; this fusion protein lacks several major structural domains of lamin B1 and therefore is assumed to be nonfunctional. The Lmnb2-knockout allele (Lmnb2–) is a null allele created by replacing exon 1 of Lmnb2 with a lacZ reporter (Coffinier et al., 2010a ). No phenotypic differences were observed between Lmnb1+/+ and Lmnb1+/Δ mice, or between Lmnb2+/+ and Lmnb2+/– mice (Vergnes et al., 2004 ; Coffinier et al., 2010a ). Conditional knockout alleles for Lmnb1 (Lmnb1fl) and Lmnb2 (Lmnb2fl) are described elsewhere (Yang et al., 2011 ). In both alleles, loxP sites were introduced flanking exon 2; Cre-mediated recombination deletes exon 2 and introduces a frameshift yielding a null mutation. Emx1-Cre transgenic mice (B6.129S2-Emx1tm1(cre)Krj/J; Gorski et al., 2002 ) and ROSA-lacZ mice (B6.129S4-Gt(ROSA)26Sortm1Sor/J; Soriano, 1999 ) were purchased from the Jackson Laboratory (Bar Harbor, ME). Mice with a Lmna lacZ reporter allele were created with a mouse embryonic stem cell line (S22-2D1) that harbored a retroviral βgeo gene-trap cassette in intron 1 of Lmna (MMRRC). Lmnb1fl/fl, Emx1-Cre Lmnb1fl/+, Lmnb2fl/fl, and Emx1-Cre Lmnb2fl/+ mice were phenotypically normal. This study was carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All animal protocols were reviewed and approved by the Animal Research Committee at the University of California, Los Angeles.
The Lmnb1Δ allele was identified by amplifying a 400–base pair product with forward primer 5′-TCC GTG TCG TGT GGT AGG AGG-3′ and reverse primer 5′-CGG AGC GGA TCT CAA ACT CTC-3′ (located within the βgeo cassette); the wild-type (WT) allele was identified by amplifying a 600–base pair product with the same forward primer and reverse primer, 5′-GCA GGA GGG TTG GGA AAG CC-3′. The Lmnb2 knockout allele (Lmnb2–) was identified by amplifying a 550–base pair product with forward primer 5′-CGG GTT TTA CTG GAA AGC TG-3′ and reverse primer 5′-GAC AGT ATC GGC CTC AGG AA-3′ (located within the lacZ cassette); the WT allele was identified by amplifying a 350–base pair product with the same forward primer and reverse primer, 5′-CGG AGC AGC AAC CTA TCA TT-3′. The Lmnb1fl allele was identified by amplifying a 450–base pair product with forward primer 5′-AAC AAA CTT GGC CTC ACC AG-3′ (located on the distal loxP site and adjacent vector sequences) and reverse primer 5′-CTG TGG GAC AAA GAC CCA GT-3′; the WT allele was identified by amplifying a 300–base pair product with forward primer 5′-CCA CTT AGC TCG GGG AGT TT-3′ and the same reverse primer. The Lmnb2fl allele was identified by amplifying a 480–base pair product with the forward primer 5′-AAC AAA CTT GGC CTC ACC AG-3′ (located on the distal loxP site and adjacent vector sequences) and reverse primer 5′-GGT CTT GAT GCC ACT CAC CT–-3′; the WT allele was identified by amplifying a 350–base pair product with forward primer 5′-TGA GGC TTT GGA GAA AAG GA-3′ and the same reverse primer. The Cre transgene was detected by PCR with the primers 5′-GCA TTA CCG GTC GAT GCA ACG AGT GAT GAG-3′ and 5′-GAG TGA ACG AAC CTG GTC GAA ATC AGT GCG-3′. Mice heterozygous for the Lmna lacZ allele were identified by amplifying the βgeo cassette with primers 5′-GAC AGT CGT TTG CCG TCT GAA TTT G-3′ and 5′-TAC CAC AGC GGA TGG TTC GGA TAA T-3′.
Paraffin-embedded sections of mouse embryos (5 μm thick) were stained with hematoxylin and eosin. For immunochemical studies, mouse tissues were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 2 h at room temperature, incubated in 30% sucrose in PBS at 4°C overnight, and frozen in O.C.T. (Tissue-Tek, Sakura Finetek). Sections 10 μm thick were fixed for 5 min in ice-cold acetone or methanol, followed by five dips in acetone and then permeabilized with 0.1% Tween-20. Background staining for mouse antibodies was minimized with the Mouse-on-Mouse Kit (Vector Laboratories, Burlingame, CA). To detect BrdU labeling, sections were pretreated with 1 N HCl for 10 min on ice, 2 N HCl for 10 min at room temperature, followed by 10 min at 37°C, and 0.1 M sodium borate (pH 8.5) for 12 min. Tissue sections were blocked with 2.5% horse serum for 1 h at room temperature and incubated overnight at 4°C with primary antibodies at the dilutions indicated in Supplemental Table S1. Alexa Fluor 488– and Alexa Fluor 568–conjugated secondary antibodies (Molecular Probes, Invitrogen, Carlsbad, CA) were used at a 1:200 dilution, and Alexa Fluor 555–conjugated streptavidin (Molecular Probes) was diluted at 20 μg/ml. Costaining of sections with two primary antibodies from the same species was accomplished by directly labeling one of the antibodies with Alexa Fluor 555 or Alexa Fluor 647 (Molecular Probes). After counterstaining with 4′,6-diamidino-2-phenylindole (DAPI), sections were mounted with Prolong Gold antifade (Invitrogen), and images were recorded with the Axiovision software on an Axiovert 200M microscope using 5× (0.16 numerical aperture [NA], EC Plan Neofluar), 10× (0.45 NA, Plan Apochromat), 20× (0.8 NA Plan Apochromat), or 40× (0.75 NA, EC Plan Neofluar) objectives with an AxioCam MRm and an ApoTome (all from Zeiss, Thornwood, NY) or with the LCS software on a TCS-SP MP laser-scanning confocal microscope with a 63× (1.4 NA, oil) objective (all from Leica, Wetzlar, Germany).
Embryos were fixed for 1 h in 4% paraformaldehyde in PBS, incubated in 30% sucrose at 4°C overnight, and embedded in O.C.T. Sections of 10 μm were cut, postfixed in paraformaldehyde, washed twice for 10 min in ice-cold PBS, and then stained for 4–16 h at 37°C in X-gal buffer (PBS, 20 mM potassium ferricyanate, 20 mM potassium ferrocyanate, 2 mM MgCl2, 0.2% NP-40, 0.1% sodium deoxycholate, and 0.8 mg/ml 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside; Nagy et al., 2003 ). Specificity of β-galactosidase staining was assessed by parallel experiments with Lmnb1+/+ tissues (which did not express β-galactosidase). After staining, sections were washed in PBS, postfixed in 4% paraformaldehyde in PBS, counterstained with eosin, dehydrated, and mounted in Permount (Fisher Scientific, Waltham, MA). Images were recorded on the Leica Application Suite imaging software with a MZ6 dissecting microscope and a DFC290 digital camera (all from Leica).
Neuronal progenitors were isolated from E13.5 sibling embryos. Cortical explants were isolated in PBS on ice and then incubated in 0.25% trypsin-EDTA (Life Technologies, Carlsbad, CA) at room temperature for 10 min; cells were further dissociated by pipetting before adding 1.5 volumes of DMEM containing 10% fetal bovine serum. Cells were plated on poly-l-lysine–coated coverslips (80,000 cells/cm2) and cultured for 4 d in neuronal differentiation medium (50% DMEM/F12-50% Neurobasal medium with B-27 and N2 supplements, all from Life Technologies). For immunocytochemistry experiments, cells were washed in PBS containing 1 mM Ca2+ and 1 mM Mg2+ (PBS/Ca/Mg) and fixed in methanol for 10 min. After permeabilizing of cells with 0.2% Triton for 5 min and blocking for 1 h in PBS/Ca/Mg containing 10% fetal bovine serum and 0.2% bovine serum albumin, the cells were incubated for 1 h with primary antibodies (Supplemental Table S1). Alexa Fluor 488–, 568–, and 647–conjugated secondary antibodies (Molecular Probes) were diluted 1:400. Nuclear DNA was stained with 2 μg/ml DAPI, and coverslips were mounted with Prolong Gold antifade reagent. Immunofluorescence images were recorded with the Axiovision software on an Axiovert 200M microscope with a 40× (0.75 NA, EC Plan Neofluar) or 63× (1.4 NA, Plan Apochromat Oil) objectives with an AxioCam MRm and ApoTome (all from Zeiss). Confocal images were recorded on the same microscope equipped with a LSM 700 laser-scanning system using the acquisition software ZEN 2010 (all from Zeiss) or on a Leica TCS-SP MP laser-scanning confocal microscope with a 63× (1.4 NA, oil) objective and the Leica Application Suite.
For all studies, at least three embryos per group were analyzed with sets of sections matched for the position in the brain. Brain sections stained for the markers of interest were photographed with a 20× objective using the Axiovision software, and numbers of cells positive for the different markers were recorded with the Count tool in Photoshop CS5 Extended (Adobe, San Jose, CA). Cells were counted in at least three areas of 430 × 350 μm taken from at least two sections. The thickness of the cortical plate and cortex were measured on images with the Ruler tool in Adobe Photoshop CS. Numbers of cells in brain sections of lamin B1–deficient and control embryos with nuclear blebs and an asymmetric distribution of lamin B2 were analyzed on three-dimensional (3D) reconstructions generated from stacks of confocal images with the 3D Opacity renderer in Volocity 5.4 (PerkinElmer, Waltham, MA). The 3D images were exported as TIF files, and counts were recorded with the Count tool in Adobe Photoshop CS5 Extended. Lengths of nuclei and distance of the nucleus to the centrosome in lamin B2–deficient cells and control cells were measured on 3D reconstructions from stacks of confocal images with Volocity 5.4 software.
Statistical analyses were performed in Excel 2004 (Microsoft, Redmond, WA). Results were expressed as mean ± SD or as a 95% confidence interval, and significant differences were analyzed with a two-tailed Student's t-test for independent samples at VassarStats (http://faculty.vassar.edu/lowry/VassarStats.html). The boxplot analysis of individual nuclear lengths was made in Stata 11 (StataCorp, College Station, TX).
This work was supported by National Institutes of Health Grants AR050200, HL76839, HL86683, HL89781, and GM66152, a March of Dimes Grant 6-FY2007-1012, and an Ellison Medical Foundation Senior Scholar Award. C.C. is the recipient of a Scientist Development Grant from the American Heart Association (0835489N). Confocal laser scanning microscopy was performed at the California NanoSystems Institute Advanced Light Microscopy/Spectroscopy Shared Resource Facility at the University of California, Los Angeles, supported with a National Institutes of Health/National Center for Research Resources shared resources grant (CJX1-44385-WS-29646) and a National Science Foundation Major Research Instrumentation grant (CHE-0722519). We thank the University of California, Los Angeles, Translational Pathology Core Laboratory and the Brain Research Institute Microscopic Techniques Laboratory Core for the preparation and sectioning of paraffin-embedded tissues.
This article was published online ahead of print in MBoC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E11-06-0504) on October 5, 2011.