Loss of Coiled-Coil Protein Cep55 Impairs Neural Stem Cell Abscission and Results in p53-Dependent Apoptosis in Developing Cortex

Abstract

To build the brain, embryonic neural stem cells (NSCs) tightly regulate their cell divisions, undergoing a polarized form of cytokinesis that is poorly understood. Cytokinetic abscission is mediated by the midbody to sever the daughter cells at the apical membrane. In cell lines, the coiled-coil protein Cep55 was reported to be required for abscission. Mutations of Cep55 in humans cause a variety of cortical malformations. However, its role in the specialized divisions of NSCs is unclear. Here, we elucidate the roles of Cep55 in abscission and brain development. KO of Cep55 in mice causes abscission defects in neural and non-neural cell types, and postnatal lethality. The brain is disproportionately affected, with severe microcephaly at birth. Quantitative analyses of abscission in fixed and live cortical NSCs show that Cep55 acts to increase the speed and success rate of abscission, by facilitating ESCRT recruitment and timely microtubule disassembly. However, most NSCs complete abscission successfully in the absence of Cep55 Those that fail show a tissue-specific response: binucleate NSCs and neurons elevate p53, but binucleate fibroblasts do not. This leads to massive apoptosis in the brain, but not other tissues. Double KO of both p53 and Cep55 blocks apoptosis but only partially rescues Cep55^-/- brain size. This may be because of the persistent NSC cell division defects and p53-independent premature cell cycle exit. This work adds to emerging evidence that abscission regulation and error tolerance vary by cell type and are especially crucial in neural stem cells as they build the brain.SIGNIFICANCE STATEMENT During brain growth, embryonic neural stem cells (NSCs) must divide many times. In the last step of cell division, the daughter cell severs its connection to the mother stem cell, a process called abscission. The protein Cep55 is thought to be essential for recruiting proteins to the mother-daughter cell connection to complete abscission. We find that Cep55 mutants have very small brains with disturbed structure, but almost normal size bodies. NSC abscission can occur, but it is slower than normal, and failures are increased. Furthermore, NSCs that do fail abscission activate a signal for programmed cell death, whereas non-neural cells do not. Blocking this signal only partly restores brain growth, showing that regulation of abscission is crucial for brain development.

Keywords: Cep55; abscission; cortical development; microcephaly; midbody; neural stem cell.

Figure 1.

Cep55 KO causes microcephaly with severe thinning of neuronal and axon layers. A, Schematics of Cep55 mouse gene and protein (not to scale). The 601 bp deletion includes all of exon 6 (311 bp) and portions of flanking introns. Cep55 protein domains include two coiled-coil regions (CC1 and CC2) surrounding the ESCRT- and Alix-binding region (EABR), and two ubiquitin binding domains (UBD). The deletion of exon 6 is predicted to result in a frameshift (FS) at amino acid 226, which causes multiple premature stop codons starting 12 residues later. B, Immunoblots of E14.5 MEF lysates show the expected Cep55 protein product at ∼55 kDa in WT and heterozygote samples, but not in KO samples. C, D, Newborn Cep55^−/− pups have 10% shorter body length and a flatter head (arrow) but otherwise appear morphologically similar to controls. E–G, Representative images and measurements show that mean cortical area is reduced in −/− brains at P0 and E14.5 compared with +/+ and +/− controls. H, Mean cortical length is significantly reduced in −/− mice at E12.5, E14.5, and P0. I, The cortical length in −/− P0 pups is disproportionately small relative to body size. J, Representative images of P0 +/− and −/− coronal sections at level of corpus callosum, stained with H&E. −/− mice have reduced cell density in neuronal layers (J′, arrow). K, Mean total cortical thickness is significantly reduced in −/− brains at P0. The vz/svz and mz thicknesses are unaltered, but the iz and cp are significantly thinner. P0 (C–I), n = 7 +/+, 14 +/−, and 8 −/− mice. E14.5 (E–H), n = 5 +/+, 13 +/−, and 10 −/− mice. E12.5 (H), n = 17 +/+;+/− and 9 −/− mice. K, n = 3 +/+, 5 +/−, and 5 −/− mice. Scale bars: C, 5 mm; E, F, 1 mm; J, 250 µm, J′, 100 µm. **p < 0.01. ***p < 0.001. ****p < 0.0001. D, G, I, One-way ANOVA; H, K, Student's t test.

Figure 2.

Disruption of cortical and cerebellar structure in Cep55 KO brains. A, Representative images of +/+ and −/− cortical sections labeled with antibodies to Ctip2, marking layers 6 (faint staining) and 5 (bright staining), and Satb2, marking layers 2-4. B, All neuronal layers are significantly thinner in −/− cortices at P0. C, Layers 5-6 occupy a larger proportion of cp, whereas layers 2-4 occupy a decreased proportion of the cp in −/− cortices. D, The absolute number of layer 6 cells (no. of nuclei/length of cortex) is reduced by 40% in −/− cortices. E, −/− layer 6 neurons are 25% less dense than WT cells (no. of nuclei/area). B, C, n = 5 +/+ and 4 −/− mice. D, E, n = 3 +/+ and 3 −/− mice. F, H, Representative Cep55^+/− and −/− sagittal (F) and coronal (H) sections stained with H&E. F, Arrows indicate approximate locations of rostral, middle, and caudal, coronal sections in C. mid, Midbrain; hind, hindbrain; PB, Probst bundles of axons; hip, hippocampus; cc, corpus callosum; dTh, dorsal Thalamus. G, The mean cortical thickness in −/− brains is 82% of normal in rostral and middle sections, and only 69% of normal in caudal sections. G, n = 3 +/+, 5 +/−, and 5 −/− mice. I-J′, Representative control and −/− sagittal sections of P0 (I) and P4 (J,J′) brains. ctx, Cortex; cbm, cerebellum. K, L, Cortex thickness, cbm length, and cbm area are reduced in −/− mice at P0 and P4. P0, n = 2 +/+, 1 +/−, and 3 −/− pups; P4, n = 1 +/+, 1 +/−, and 2 −/− pups. Scale bars: A, 100 µm; F, H, I, 500 µm; J, J′, 1 mm. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. F, One-way ANOVA. others: t test.

Figure 3.

Cep55 KO cortex shows reduced and disorganized neuron and NSC layers by E14.5. A, Representative images of cortical sections from Cep55 +/− and −/− embryos immunostained for neuronal tubulin (Tubb3, gray) show cp and axons in the iz. In −/− images, the cp is disorganized. B, Mean thicknesses of total cortex and iz are significantly decreased in −/− brains. C, The −/− cp/iz occupies proportionally less of cortical width than normal, while the vz/svz occupies more. D, Representative images of E14.5 Cep55^+/− and −/− coronal, middle sections stained with H&E. E, The iz and cp are ∼ 30% thinner in −/− cortices compared with controls; however, the mz is 20% thicker. F, Cortical sections stained for NSC marker Pax6 (red) and BP marker Tbr2 (green) show that NSCs in the vz of Cep55^−/− brains are more disorganized, with some empty spaces (square), and some NSC nuclei mislocalized above the svz (arrows). G, H, NSCs (Pax6⁺) per cortical length are reduced in −/−, and some are mislocalized. BP numbers (Tbr2⁺) are not significantly changed. I, PH3 immunostaining marks cells in mitosis. J, −/− cortices show a normal number of mitotic cells at the apical membrane but an increased number of basal mitotic cells. K, The mitotic index of NSCs (Pax6⁺) is normal, but of BPs (Tbr2⁺) is increased in −/− cortices. L, Representative images of E12.5 control (+/+, +/−) and −/− coronal cortical sections immunostained for neuronal tubulin (Tubb3, white) to label preplate (pp) and BP marker Tbr2 (green). M, The mean cortical thickness is not different between control and −/− brains. N, O, Preplate of −/− brains is not thicker but is an increased proportion of total thickness. P, Cep55^−/− cortices have decreased Tbr2 cell number per 100 µm of the VZ. A, F, I, L, Dashed line indicates apical membrane. B, Total thickness n = 7 +/− and 8 −/− brains. B, Layer thicknesses. C, F-K, n = 4 +/− and 4 −/− brains. E, n = 6 +/− and 5 −/− mice. M-O, n = 6 controls (4 +/+, 2 +/−), 6 −/− brains. P, n = 3 controls (2 +/+, 1 +/−), 3 −/− brains. pp, Preplate. Scale bars: A, L, 20 µm; D, 100 µm. *p < 0.05; **p < 0.01; ***p < 0.001; t test.

Figure 4.

Cep55 protein localizes in late-stage midbodies of NSCs, BPs, and MEFs. A, B, Representative images of Cep55^+/− control and Cep55^−/− E14.5 NSCs undergoing cytokinesis that were cultured for 1 day in vitro (DIV) fixed and immunostained for endogenous Cep55 (red), Aurora kinase B (green), Nestin for NSCs (gray), and DAPI. Cep55 is not detectable in mitotic spindles or centrosomes in metaphase, anaphase, or early midbody stage NSCs but accumulates in the central bulge of late midbodies and in MBRs in control (A) but is undetectable in KO (B) NSCs. C, Quantification of endogenous Cep55 immunofluorescence signal intensity in control and −/− NSC midbodies. D, Cep55 in a late-stage midbody of a BP from control brain (left). Cep55 does not localize to midbodies of BPs from KO brains (right). E, Cep55 signal was not detected at centrosomes, marked by pericentrin, in NSCs (open arrows), but is detected in a nearby MBR (closed arrow). F, G, Cep55^+/+ MEFs show a similar pattern of Cep55 localization as NSCs, with accumulation only in late-stage midbodies (F), but is undetectable in −/− MEF midbodies (G). Scale bars: A, metaphase and early midbody: 5 µm; D, 4 µm; E, 5 µm; F, 10 µm; inset, 5 µm. ****p <0.0001, t test.

Figure 5.

Cep55 KOs display NSC midbody defects in fixed cortical slab preparations. A, B, Schematics of cortical slab dissection, cross section versus apical membrane views of NSCs undergoing cytokinesis. Images of midbodies (MB) at different stages of the abscission process: early (wider), late-stage (thinner), and an MBR. C, Representative images of E14.5 cortical slabs immunostained for apical junctions (ZO-1), mitotic chromatin (PH3⁺), or MBs (AurkB +, arrows). D, Mean apical endfoot density is reduced in −/− slabs. E12.5: n = 4 +/− slabs (4 brains), 4 −/− slabs (4 brains). E14.5: n = 8 control slabs (6+/+, 2+/− brains) and 7 −/− slabs (6 brains). E, F, Mitotic index is normal, but MB index is significantly increased. E, n = 6 control slabs (4 +/+, 2 +/− brains), 6 −/− slabs (5 brains). F, E12.5: n = 4 +/− slabs (4 brains), 4 −/− slabs (4 brains). E14.5: n = 6 control slabs (4 +/+, 2 +/− brains), 5 −/− slabs (5 brains). G, A smaller percentage of −/− MBs have visible cs. H, −/− MBs tend to be shorter than controls (measured by AurkB). Medians: 2.7 µm for +/+, 2.5 µm for −/−. Bin = 0.3 µm. G, H, n = 353 control MBs (5 +/+, 2 +/− brains), 246 −/− MBs (5 brains). I, Cortical slabs immunostained for AurkB to mark pre-abscission MB flanks (brackets), and CitK to mark MB bulges and post-abscission MBRs (arrowheads). J, K, MBRs are increased in −/− brains, normalized to MB number or NSC (endfoot) number. J, K, E12.5 and E14.5: n = 4 control slabs (2 +/+, 2 +/− brains), 4 −/− slabs (4 brains). Scale bars: B, 1 µm; C, 2 µm; I, 5 µm. *p < 0.05. **p < 0.01. ****p < 0.0001. D, F, G, t test, J, K, ANOVA; E, Fisher's exact test.

Figure 6.

Cep55 KO causes delayed midbody microtubule disassembly during NSC abscission. A, Schematics and time-lapse images of an E13.5 NSC in a cortical slab explant undergoing midbody (MB) abscission. Microtubules are labeled by SiR-Tubulin. Distinct steps shown are as follows: MB formation, flank 1 thinning, microtubule disassembly on flank 1 (first abscission), and microtubule disassembly on flank 2 (second abscission). After bilateral flank disassembly, the MBR is left at the apical membrane. B–D, Time from MB formation to complete first abscission and second abscission are increased in Cep55^−/− NSCs, but there is no change in the percentage of bilateral abscissions detected. E, F, Cumulative frequency plots for the first abscission and second abscission show the curves are shifted to the right in −/− NSCs. G, Control NSC with a bipolar spindle (asterisks at spindle poles) proceeded to form a bipolar midbody at 60 min. Cep55^−/− NSC with a tripolar spindle (asterisk at spindle poles) proceeded to form a tripolar midbody at 105 min (arrowheads point to midbody “flanks”). H, Tripolar spindles are increased in −/− cortices. B, C, E, n = 54 control cells (2 +/+, 1 +/− slabs), 46 −/− cells (4 slabs). D, F, n = 43 control cells, 39 −/− cells. G, n = 71 control cells (2 +/+, 1 +/− slabs) and 86 −/− cells (4 slabs). Scale bars: A, 1 µm; G, 10 µm. *p < 0.05. **p < 0.01. B, D, t test. E, F, Kolmogorov–Smirnov and Mann–Whitney tests. C, H, Fisher's exact test.

Figure 7.

Cep55 KO NSCs and MEFs have reduced, but not eliminated, ESCRT recruitment to midbodies. A, Model from the literature for Cep55 recruiting ESCRTs in midbodies (MBs). B, Representative images of MBs of NSCs, from control or Cep55^−/− E14.5 embryos, immunostained for ALIX or TSG101. C, −/− NSC MBs are much less likely to have detectable Alix or TSG101. Alix: n = 49 NSC control MBs (3 animals, 3 coverslips); n = 46 −/− MBs (2 animals, 3 coverslips); TSG101: n = 32 control (2 animals, 2 coverslips), n = 45 −/− (3 animals, 3 coverslips); all, 1 DIV. D, Representative images of MBs of 1 DIV MEFs, from control or −/− E14.5 embryos, immunostained for ALIX, TSG101, or Chmp2A. Arrowhead indicates constriction site (CS). E, Central dark zones were detected at normal percentage in −/− MEF MBs, using α-tubulin or AurkB immunostaining. n = 402 control, 355 −/− MBs (5 animals, 6 coverslips each). F, Cep55^−/− MEF midbodies are less likely to have detectable Alix, TSG101, or Chmp2a than controls. Alix: n = 153 control, 129 −/− MBs (4 animals, 4 coverslips each); Tsg101: n = 141 control, 132 −/− MBs (4 animals, 4 coverslips each); Chmp2a: n = 108 control, 105 −/− MBs (3 animals, 3 coverslips each). G, The % of MBs with CSs that contain detectable ESCRT is decreased in −/− MBs. n (MBs with CSs) = 110 control, 95 −/− (5 animals, 11 coverslips each). H, I, MBs with ESCRT present were analyzed by fluorescence intensity line scans through the central dark zones. Alix peak is severely reduced in −/− MBs, but Tsg101 and Chmp2a peaks are not. Alix: n = 11 control, −/− MBs (2 animals, 2 coverslips, 1 experiment); Tsg101: n = 14 control, 15 −/−MBs (2 animals, 2 coverslips, 1 experiment); Chmp2a: n = 15 control, −/− MBs (2 animals, 2 coverslips, 1 experiment). Scale bars: B, D, 1 µm. *p < 0.05; **p < 0.01; ***p < 0.001; C-G, Fisher's exact test; H-J, t test.

Figure 8.

Cep55 KO mice have increased numbers of binucleate cortical cells and MEFs. A, Flow cytometric analysis of E15.5 dissociated cortices labeled with propidium iodide (PI) indicates an increase in cells with tetraploid (4N) DNA content in Cep55^−/− brains, and concomitant decrease in cells with 2N DNA content. B, C, Both proliferating (Ki67⁺) and nonproliferating (Ki67^–, presumed neurons) populations of −/− brains show increases in 4N DNA cells. D–F, Increased numbers of binucleate progenitors (Nestin⁺) and neurons (Nestin^–) are seen in Cep55^−/− dissociated cortical cultures at 1 DIV. D, Arrow indicates example of mononucleate progenitor. E, Arrow indicates example of binucleate neuron. G, H, Primary cultures of MEFs from −/− embryos have twice as many binucleate cells as control cultures at 1 DIV. Arrow indicates example of mononucleate cell. Double arrow indicates example of binucleate cell. A, n = 6 Cep55^+/+;+/−, 4 −/− dissociated cortices. B, C, n = 5 +/+;+/−, 5 −/− dissociated cortices. F, n = 4 +/+;+/− and 4 −/− coverslips from 2 embryos each. H, n = 5 +/−, 5 −/− coverslips from 3 embryos each. Scale bars: D, G, 10 µm. **p < 0.01. ***p < 0.001. A, F, H, t test.

Figure 9.

Apoptosis is greatly increased in embryonic neural tissue of Cep55 KO embryos. A, Cortical sections immunostained for apoptotic cell marker CC3 (green) show almost no CC3⁺ cells in control but many in Cep55^−/− cortex. B, At E14.5, apoptotic cells were most increased in the vz/svz, containing nuclei of NSCs and BPs, but also increased in neurons in cp/iz of −/− brains. C, D, At E10.5, before neurogenesis, apoptosis is greatly increased in −/− forebrain NSCs. E-E′′, Representative images of coronal cortical sections from E14.5 Cep55^−/− brains immunostained for neuron markers Satb2 (green) and Ctip2 (red) and apoptotic marker CC3 (gray). E′, Inset of top box in A shows a CC3⁺ Ctip2⁺ cell. E′′, Inset of bottom box in A shows a CC3⁺ Satb2⁺ cell. Arrows indicate additional CC3⁺ Ctip2⁺ cells. Arrowheads indicate additional CC3⁺ Satb2⁺ cells. F, There are increased numbers of CC3⁺ cells (cells defined as DAPI⁺) in Cep55^−/− cortex at E12.5 compared with E14.5. G, The proportion of CC3⁺ cells that are copositive for neuron marker Tubb3 is slightly increased at E14.5 compared with E12.5. H, A similar percentage of Satb2⁺ and Ctip2⁺ cells are copositive for CC3 at E14.5. I-J′, Representative images of cortical sections from E12.5 (I) and E14.5 (J) −/− cortex immunostained for BPs (Tbr2) and CC3. Copositive cells are observed (arrows and box inset, J′). K, The proportion of CC3⁺ cells that are copositive for Tbr2 is increased at E14.5 compared with E12.5. L, Approximately 10% of Tbr2⁺ cells are copositive for CC3 at E12.5 compared with 8% at E14.5 in −/− cortices. M, N, Apoptosis is high in the E10.5 −/− spinal cord, (and midbrain and hindbrain, not shown), but not in heart, forelimb, or retina. O, P, The brain specificity of the apoptosis is not simply because of higher mitotic index. A, C, Dashed line indicates apical membrane. All experiments, n = 3 control and −/− brains or embryos at each age. Scale bars: A, C, E, 20 µm; E′, J′, 5 μm; I, 10 μm; M, left, 40 µm; retina, 20 µm. *p <0.05. **p < 0.01. ***p < 0.001. All experiments, t test.

Figure 10.

p53 nuclear accumulation is increased in Cep55 KO binucleate NSCs and neurons, but not in binucleate MEFs. A, B, Cortical sections immunostained for p53 and Tubb3 show almost no p53⁺ cells in controls but many in Cep55^−/− cortices. Arrows indicate paired nuclei with p53 expression. B, p53⁺ cell counts are increased in Cep55^−/− cortices at E10.5, and at E14.5, primarily in the vz/svz. C, Images of dissociated NSCs cultured 1 DIV show p53⁺ binucleate NSCs from −/− cortices. D, Cells with a p53 N:C ratio of >2 are greatly increased in −/− cultures, both NSCs and neurons. E, In −/− NSC cultures, the mean N:C ratio of p53 intensity is ∼1 in mononucleate NSCs or neurons, but significantly higher in binucleate progenitors and neurons. F, In −/− cultures, over half of binucleate NSCs have a p53 N:C ratio >2, compared with only 1% of mononucleate NSCs. Among binucleate neurons, ∼20% have a p53 N:C ratio of >2 versus only 2% of mononucleate neurons. H, Apoptosis (CC3⁺) is not increased in −/− primary MEFs cultured 1 DIV. I, The N:C ratio of p53 signal (p53, green, G) is not different in −/− MEFs compared with controls. J, Binucleate −/− MEFs (G, arrows) did not have an increased p53 N:C ratio compared with mononucleates. A, Dashed line indicates apical membrane. B, n = 3 +/+;+/− and 3 −/− mice at each age. D-F, n = 4 +/+ or +/− and 4 −/− coverslips from 2 embryos each; 124 control and 137 −/− NSCs; 143 −/− and 184 −/− neurons for N:C ratios. H, n = 3 control and 3 −/− coverslips from 3 embryos each. I, n = 195 control and 232 −/− cells from 3 control and −/− coverslips and embryos each. J, n = 153 mononucleate and 25 binucleate control cells and 158 mononucleate and 68 binucleate −/− cells from 3 coverslips and 3 embryos each. Scale bars: A, C, G, 20 µm. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. All experiments, t test.

Figure 11.

p53 KO blocks apoptosis and partially rescues brain size but exacerbates progenitor defects of Cep55 KO. A–C, Codeletion of p53 partially rescues microcephaly in Cep55^−/− mice. B, C, Homozygous p53 KO shows partial rescue of Cep55^−/− cortical length and area. D, E, Apoptosis is fully blocked in Cep55^−/−;p53^−/− dKOs. F, Representative images of E14.5 cortical sections from control (Cep55^+/−;p53^−/−) and dKOs immunostained for neuron marker Tubb3 (gray) show cp and axons in the iz. In dKO images, there is disorganization of the cp and iz boundary. G, Mean thicknesses of total cortex and vz are slightly decreased in dKO brains. H, The dKO cp occupies proportionally more of cortical width than in controls. I, Cortical sections stained for NSC marker Pax6 (red) and BP marker Tbr2 (green) show NSCs and BPs in the vz and svz of dKO brains are more disorganized, with many nuclei mislocalized basally above the svz (arrows), and some large nuclei (square). J, K, There is no difference in the number of total or localized NSCs (Pax6⁺) or BPs (Tbr2⁺) per cortical length in dKOs. L, PH3 (magenta) immunostaining is used to mark cells in mitosis. Mislocalized mitotic cells are noted in dKO cortices (arrows). M, dKO cortices show increased numbers of mitotic cells, with a normal number of mitotic cells at the apical membrane but an increased number basally. N, The mitotic index of both NSCs (Pax6⁺) and BPs (Tbr2⁺) is significantly increased in dKO cortices. O, There are increased numbers of mitotic NSCs mislocalized above the apical membrane, and mitotic BPs mislocalized above the svz, in dKO sections. D, F, I, L, Dashed line indicates apical membrane. B, C, n = 11 controls (Cep55^+/+,+/−;p53^+/+,+/−), 2 Cep55^−/−;p53^+/+, 5 Cep55^−/−;p53^+/−, 3 Cep55^−/−;p53^−/−. E, n = 3 controls (Cep55^+/+,+/−;p53^+/+,+/−), 2 Cep55^−/−;p53^+/+, 3 Cep55^−/−;p53^−/−. F-O, n = 5 control (Cep55^+/+,+/−;p53^−/−) and 5 Cep55^−/−; p53^−/− brains. Scale bars: A, 0.5 mm; D, 40 µm; F, I, L, 20 µm. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. B, C, E, One-way ANOVA; others, t test.

Figure 12.

Cell cycle exit is increased in Cep55^−/− cortices at E12.5 and is p53-independent. A, Representative images of coronal cortical sections from Cep55 control (+/+, +/−) and −/− embryos injected with BrdU at E11.5 and collected at 24 h, immunostained for neuron marker Tubb3 (gray), BP marker Tbr2 (green), and recently divided cells (BrdU⁺ only, red). B, Increased numbers of BrdU⁺ nuclei were copositive for Tubb3⁺ in Cep55^−/− cortices. n = 308 control, 204 −/− nuclei from 3 control and 3 −/− brains. C, Schematic of modified BrdU pair cell assay. E12.5 progenitors were dissociated and plated with BrdU followed by washout, incubation, fixation, and immunostaining with BrdU, Tubb3, and Nestin (progenitors). Pairs of BrdU⁺ cells were identified and marked as containing progenitors (Nestin⁺, Tuj1^–), or neurons (Tuj1⁺, Nestin⁺, or Nestin^–). D, Representative images of BrdU⁺ pairs that are P-P, N-N, or N-P. E, Cep55^−/− NSCs have reduced numbers of P-P and increased N-N divisions. n = 309 +/+, 256 −/− pairs; 3 coverslips/animals each. F, There are more BrdU⁺Tubb3⁺ nuclei indicating cell cycle exit in −/− cultures. G, Representative image of paired nuclei that are Tbr2⁺. n = 750 +/+ and 622 −/− nuclei; 3 coverslips/animals each. H, The percentage of Tbr2⁺ nuclei in a BrdU pair is not altered in −/− cultures. n = 566 +/+ and 528 −/− nuclei; 3 coverslips/animals each. I, dKO E12.5 NSCs have reduced proliferative symmetric and increased neurogenic symmetric division compared with controls. J, There are more BrdU⁺Tubb3⁺ nuclei in dKO cultures. K, L, n = 217 Cep55^+/+;p53^−/− pairs/528 nuclei; 3 coverslips/2 animals and 189 Cep55^−/−;p53^−/− pairs/481 nuclei; 4 coverslips/2 animals. Scale bars: A, C, 20 µm; F, I, 5 µm. *p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001. B, D, G, K, χ²; others, Fisher's exact. K, Working model for the consequences of Cep55 loss on abscission of mammalian cells. Cep55^−/− mouse cells have impaired ESCRT recruitment and abscission delay. Most cells complete abscission, but at least some fail to complete cytokinesis resulting in increased binucleate cells. Binucleate NSCs and neurons have elevated p53 expression, but not binucleate MEFs. p53 activation in NSCs and neurons causes apoptosis in Cep55 KO cortices, which partially accounts for microcephaly. Precocious cell cycle exit occurs after some NSC divisions and is p53-independent.