|Home | About | Journals | Submit | Contact Us | Français|
B-type lamins, the major components of the nuclear lamina, are believed to be essential for cell proliferation and survival. We found that mouse embryonic stem cells (ESCs) do not need any lamins for self-renewal and pluripotency. Although genome-wide lamin-B binding profiles correlate with reduced gene expression, such binding is not directly required for gene silencing in ESCs or trophectoderm cells. However, B-type lamins are required for proper organogenesis. Defects in spindle orientation in neural progenitor cells and migration of neurons probably cause brain disorganizations found in lamin-B null mice. Thus, our studies not only disprove several prevailing views of lamin-Bs but also establish a foundation for redefining the function of the nuclear lamina in the context of tissue building and homeostasis.
The major structural components of the nuclear lamina found underneath the inner nuclear membrane in metazoan nuclei are type V intermediate filament proteins called lamins (1). Mammals express both A- and B-type lamins encoded by three genes, Lmna, Lmnb1, and Lmnb2. Lmnb1 and Lmnb2 express lamin-B1 and -B2, respectively. Lmnb2 also expresses lamin-B3 through alternative splicing in testes. Mutations in lamins have been linked to a number of human diseases referred to as laminopathies (2), although the disease mechanism remains unclear. A-type lamins are expressed only in a subset of differentiated cells and are not essential for basic cell functions (3, 4). By contrast, at least one B-type lamin is found in any given cell type. Because numerous functions, including transcriptional regulation, DNA replication, and regulation of mitotic spindles, have been assigned to B-type lamins, they are thought to be essential for basic cell proliferation and survival (1, 5–8).
The nuclear lamina can regulate gene silencing by creating a repressive environment or through direct binding of lamins to genes to suppress their expression (9–13). Global mapping of chromatin regions that interact with lamin-B1 indicates that lamin-B1–associated domains generally show low gene expression (14). Moreover, interactions between lamin-B1 and chromatin change after embryonic stem cells (ESCs) differentiate, and increased lamin-B1 binding correlates with silencing of genes, which is important for lineage specification. Thus, B-type lamins have been proposed to be essential for transcriptional silencing during differentiation (15).
Although B-type lamins are reported to function in seemingly unrelated cellular processes, studies have questioned some of these roles (16–20). Published analyses have been done under conditions in which not all lamin proteins were removed. Thus, definitive tests for the functions of lamins are lacking. To study the role of lamin-B in transcription, we first mapped lamin-B1–chromatin interactions in wild-type ESCs and trophectoderm (TE) cells derived from ESCs by expressing a fusion protein of lamin-B1 with DNA adenine methyltransferase and measuring methylated regions of the genome (Dam-ID analyses) (14).We found that lamin-B1 binding patterns in the ZHBTc4 ESCs (21) that we used were similar to those of other ESCs (15), but differed from those of TE cells, fibroblasts, neural progenitor cells, and astrocytes at both large (15) and fine (0.5 to 2 kb) scales surrounding promoters (figs. S1 to S3 and table S1). Most developmentally controlled genes exhibited increased interaction with lamin-B1 after ESCs differentiated to TE cells (Fig. 1A, fig. S3D, and table S1). Lamin-B1 binding was inversely related to gene expression in ESCs based on the published RNA-seq (22) (Fig. 1B). Interactions of lamin-B1 with promoters of down-regulated genes increased after differentiation of ESCs to TE cells (Fig. 1C, left). Thus, lamin-B binding to promoters correlates with suppressed transcription.
Next, we created Lmnb1 and Lmnb2 knockout mice (fig. S4) and derived ESCs (23) with the following genotypes: Lmnb1+/+Lmnb2+/+, Lmnb1−/−Lmnb2+/+, Lmnb1+/+Lmnb2−/−, and Lmnb1−/−Lmnb2−/− from blastocysts obtained by mating Lmnb1+/−Lmnb2+/− mice. Lmnb1−/−Lmnb2−/− ESCs did not express any lamin-Bs (Fig. 2, A and B, and fig. S5A). Lamin-A and lamin-C transcripts and proteins were also not detected in all ESCs (fig. S5, A and B). Thus, the Lmnb1−/−Lmnb2−/− ESCs did not express any lamins. Surprisingly, nuclear envelopes and nuclear pores in lamin-null ESCs were indistinguishable from those of controls (Fig. 2, C and D). Although lamin-null ESCs had a slight delay in prometaphase (fig. S6), they expressed pluripotent markers, had colony morphologies (fig. S5, C and D) and growth rates (fig. S7), and maintained euploidy similar to those of controls (table S2). Thus, lamins are not essential for ESCs.
We next induced TE differentiation by expressing tamoxifen-driven Cdx2, a transcription factor required for TE development (24). In the absence of tamoxifen, Lmnb1+/+Lmnb2+/+ and Lmnb1−/−Lmnb2−/− ESCs expressed similar amounts of pluripotency markers (Fig. 3A). Cdx2 induction by tamoxifen drove TE differentiation in both wild-type and lamin-null ESCs with similar efficiencies as judged by protein or RNA abundance for TE markers (Fig. 3A and fig. S8). Microarray analyses revealed that although deleting B-type lamins resulted in changes of gene expression in ESCs or TE cells, respectively (Fig. 3, B to D, and tables S3 and S4), the changes did not correlate with lamin-B1 binding to promoters of these genes as measured in wild-type cells (Fig. 3E). Because decreased expression of genes promoting pluripotency and increased expression of TE cell-specific genes occurred before the expression of A-type lamins in TE cells (Fig. 3A), A-type lamins were unlikely to compensate for B-type lamins in regulating expression of these genes. Thus, despite the correlation between lamin-B binding and gene silencing, lamin-Bs do not appear to directly regulate expression of their bound genes in ESCs or TE cells.
Analyses of progeny from intercrosses of Lmnb1+/−Lmnb2+/− mice at embryonic day 12.5 (E12.5) and E18.5 revealed Mendelian distribution of all allelic combinations (table S5). Whole-mount histology of Lmnb1−/−Lmnb2−/− embryos at E12.5, 15.5, and 18.5 revealed all internal organs as seen in control littermates (fig. S9, A to C). However, body sizes of Lmnb1−/−Lmnb2−/− and Lmnb1−/−Lmnb2+/+ embryos appeared smaller from E14 to 15, and by E18.5 they were significantly smaller than Lmnb1+/+Lmnb2−/− and Lmnb1+/+Lmnb2+/+ mice (Fig. 4A and fig. S9, B to D). All lamin-B mutant pups were born with heartbeats and paw pinch reflexes, indicating some cardiac and muscular functions, respectively. However, they all failed to breathe and died immediately. The lungs of Lmnb1−/−Lmnb2−/− and Lmnb1−/−Lmnb2+/+ mice exhibited smaller alveoli compared to those of Lmnb1+/+Lmnb2+/+ and Lmnb1+/+Lmnb2−/− littermates (fig. S10A). The diaphragms were extremely thin with poor phrenic nerve innervations in all lamin-B mutants compared to control littermates (fig. S10, B, C, and E). Both Lmnb1−/−Lmnb2−/− and Lmnb1−/− Lmnb2+/+mice exhibited microcephaly, whereas the brain sizes of Lmnb1+/+Lmnb2−/− mice were similar to those of control littermates (Fig. 4B and fig. S10D). Thus, lamin-Bs are required for proper development of multiple organs, and defects in lungs, diaphragms, and brains may explain the failure of lamin-B mutant mice to breathe.
At E18.5, the cerebral cortices of Lmnb1−/−Lmnb2−/− brains were significantly thinner than those of control littermates (434.42 ± 20.57 versus 711.45 ± 23.76 mm, n = 9 sections for each, SEM; t test, P < 10−6). Cortical layers of brains from all lamin-B mutant mice were disorganized (Fig. 4, C and D, and figs. S11 and S12). Embryonic cerebral cortices lacking A- and B-type lamins did not exhibit large nuclear shape changes (figs. S13 and S14). Analyses of E14.5 cortices with markers for neurons (TuJ1) and subventricular progenitors (Tbr2) or E18.5 cortices with TuJ1 and Nestin (neural progenitors) indicated that the relative position and thickness of neuronal and progenitor cell layers were largely maintained in all lamin-B mutants (fig. S15). Therefore, the disorganization of cortical layers was not simply due to the lack of neuronal differentiation or abnormal expansion of progenitor cells.
Because lamin-B interacts with NudEL (7), a regulator of dynein known to function in spindle orientation, interkinetic nuclear migration (INM), and neuronal migration (25), we tested whether lamin-B is required for these cell functions. Cell cleavage planes in the neural progenitor cells (NPCs) of all lamin-B mutants were significantly more horizontal to the apical surface than those of Lmnb1+/+Lmnb2+/− mice (Fig. 4, E to G). Therefore, both lamin-B1 and -B2 are required for proper orientation of mitotic spindles in NPCs. At E18.5, an increased number of NPCs, as judged by marker staining, were scattered into basal layers in Lmnb1−/−Lmnb2−/− (fig. S16), and at E14.5,G2/MphaseNPCs, as indicated by phospho-histone-H3 label, were also present in more basal positions in Lmnb1−/−Lmnb2−/− compared to controls (fig. S17). This ectopic localization of NPCs may contribute to the enhanced cell cycle exit rates observed (Fig. 4H and fig. S18A) despite the normal cell cycle length (fig. S18B). It also suggested that the nuclei of NPCs failed to undergo INM to the apical surface. Because NPCs of Lmnb1−/−Lmnb2+/+and Lmnb1+/+Lmnb2−/− mice had normal localizations (figs. S16 and S17), lamin-Bs may have redundant functions in INM.
To study neuronal migration, we labeled neurons born at E14.5 with bromodeoxyuridine (BrdU) and analyzed their localization at E18.5. The BrdU+ neurons remained largely near the ventricular or subventricular zones (VZ/SVZ) in all lamin-B mutant brains, whereas in control brains, the BrdU+ neurons moved away from the VZ/SVZ where they were born (Fig. 4I). The BrdU+ neurons in Lmnb1+/+Lmnb2−/− mice migrated farther away from VZ/SVZ than those of Lmnb1−/−Lmnb2−/− and Lmnb1−/−Lmnb2+/+mice (Fig. 4I). Thus, lamin-B1 has a larger role in neuronal migration than lamin-B2, which could explain why Lmnb1−/−Lmnb2−/− andLmnb1−/−Lmnb2+/+mice exhibited increased apoptosis in neocortices (fig. S19) and decreased brain and body sizes than those of Lmnb1+/+Lmnb2−/−mice (Fig. 4, A and B).
The surprising lack of a requirement for lamins in ESCs and during ESC differentiation in vitro could reflect the nonengagement of these cells in tissue building and thus their lack of a need for spindle orientation and cell migration. The cellular functions of lamin-Bs we uncovered here suggest that lamin-Bs have evolved to facilitate the integration of different cell types into the highly elaborate tissue architectures found in animals. Lamin-B directly binds to NudEL to regulate spindles in vitro (5–7), yet the physiological relevance of this mitotic role has remained unclear. Our findings support both the function of lamin-Bs in dynein-mediated spindle orientation in mitotic NPCs and the role of lamin-Bs in dynein-mediated neuronal migration and INM. By refuting several well-accepted functions of lamin-Bs, this work has set the stage to dissect the mechanism by which the nuclear lamina promotes tissue organization and functions.
Note added in proof: While this work was in review, a role of lamin-B1 and -B2 in brain development has been described byC. Coffinier et al. (26).
We thank B. Van Steensel and H. Niwa for reagents, M. Sepanski for electron microscopy, O. Martin for technical support, C. Liu (National Heart, Lung, and Blood Institute) for mouse targeting, and the members of the Zheng lab for critical comments. Supported in part by the Intramural Research Program of National Institute on Aging (A.A.S. and M.S.H.K.) and R01 GM056312 (Y.Z.). Y.Z. is supported by the Howard Hughes Medical Institute. A patent (7510850) regarding the use of Xenopus egg extracts to isolate and study the lamin-B spindle matrix is held by The Carnegie Institution for Science.
Supporting Online Material
Materials and Methods
Figs. S1 to S19
Tables S1 to S7