Cep55 promotes cytokinesis of neural progenitors but is dispensable for most mammalian cell divisions

Abstract

In mammalian cell lines, the endosomal sorting complex required for transport (ESCRT)-III mediates abscission, the process that physically separates daughter cells and completes cell division. Cep55 protein is regarded as the master regulator of abscission, because it recruits ESCRT-III to the midbody (MB), the site of abscission. However, the importance of this mechanism in a mammalian organism has never been tested. Here we show that Cep55 is dispensable for mouse embryonic development and adult tissue homeostasis. Cep55-knockout offspring show microcephaly and primary neural progenitors require Cep55 and ESCRT for survival and abscission. However, Cep55 is dispensable for cell division in embryonic or adult tissues. In vitro, division of primary fibroblasts occurs without Cep55 and ESCRT-III at the midbody and is not affected by ESCRT depletion. Our work defines Cep55 as an abscission regulator only in specific tissue contexts and necessitates the re-evaluation of an alternative ESCRT-independent cell division mechanism.

Fig. 1. Cep55-knockout mice are viable.

a Schematic representation of mouse Cep55 protein domains (EABR, ESCRTs and ALIX-binding domain; MB, midbody) and Cep55 genomic locus, showing wild-type allele (+), the knockout first allele tm1a (including the selection cassette (neo), the LacZ trapping cassette, and LoxP and FRT recombination sites), the conditional allele tm1c (F, floxed), and the deletion allele tm1d (−). b PCR analysis of primary mouse tail tip fibroblasts (TTFs) with primers P1–4 shown in a to verify Cep55 status; n = 3 independent experiments. c Images of newborn (P0) mice of the indicated genotypes. Dotted lines indicate skull shape. d Body weights of mice of the indicated genotypes and developmental stages. Horizontal bars indicate mean; n = 5, 20, 5; 17, 24, and 14 mice, respectively; P-values calculated using one-way ANOVA followed by Dunnett’s multiple comparisons test. e Western blott of protein extracts from TTFs and mouse embryonic fibroblasts (MEFs) of the indicated genotypes with antibodies against Cep55 and Gapdh; n = 3 independent experiments. Cep55^−/− in b, c, d, and e indicates Cep55^tm1a/tm1a mice. Source data for e and d are provided as a Source Data file.

Fig. 2. Cep55-knockout mice show microcephaly.

a Brain images from P0 mice of the indicated genotypes. b Weights of the indicated organs and genotypes. Horizontal bars indicate mean; n = 3, 13, 5 (E18.5-brain); 11, 20, 12 (P0-brain); 4 and 4 (P0-lungs, liver, heart, and kidneys); 5, 20, and 5 (E18.5-body without (w/o) brain) mice, respectively; P-values calculated using one-way ANOVA followed by Dunnett’s multiple comparisons test where more than two sets of data are compared, otherwise Student’s two-tailed unpaired t-test. c Western blot of protein extracts from brain cortices of the indicated genotypes and developmental stages with antibodies against Cep55 and Actin; n = 3 independent experiments. d Sagittal sections of P0 brains stained with hematoxylin and eosin (HE). Dotted black lines indicate cortical dimensions measured in e (curved) and f (straight). Dotted red boxes indicate area enlarged in g. e, f Quantification of cortical length (e) and cortical thickness (f) at the indicated developmental stages. Control includes Cep55^+/+ and Cep55^+/− mice. n = 3 mice per genotype. g Enlarged view of the forebrain cortices from d. CP, cortical plate; IZ, intermediate zone; VZ, ventricular zone. h Cell counts in a 200 µm-wide field of neocortex in control and Cep55^−/− mice. n = 3 mice per genotype. e, f, h Horizontal bars indicate mean; error bars indicate SEM; P-values calculated using Student’s two-tailed unpaired t-test. a–g Cep55^−/− indicates Cep55^tm1a/tm1a mice (see Fig. 1a for allele details). Source data for b, c, e, f, h are provided as a Source Data file.

Fig. 3. Cep55 deletion results in apoptotic cells and a reduction of neural progenitors in the brain cortex.

a HE staining of comparable regions of Cep55^+/+ and Cep55^−/− cortices at E16.5. Boxed areas are magnified in b. Black arrow, single pyknotic nucleus; red arrow, pyknotic doublet. c Quantification of pyknotic nuclei in embryonic Cep55^+/+ and Cep55^−/− cortices. n = 3 mice per genotype; 4039, 2330, 2189, and 1067 cells quantified, respectively. d, e Immunohistochemical staining of E13.5 (d) and E16.5 (e) cortices for active Caspase 3 (C3A). Boxed areas in e are magnified in f. Black arrow, single C3A-positive cell; red arrow, C3A-positive doublet. g Quantification of C3A-positive (C3A+) cells in embryonic Cep55^+/+ and Cep55^−/− cortices. n = 3 mice per genotype; 4076, 2290, 2087, and 1123 cells quantified respectively. h Quantification of C3A+ cells in embryonic Cep55^−/− cortical layers. n = 3 mice per genotype; 1123 and 2290 cells quantified, respectively. i Immunohistochemistry of Cep55^+/+ and Cep55^−/− mouse cortices for the apical neural progenitor marker Pax6. j Quantification of the thickness of the Pax6-positive ventricular zone (Pax6+ thickness) (left) and the percentage of Pax6+ nuclei (right) in cortical sections like those in i. n = 3 mice per genotype at E13.5; 2 mice per genotype at E16.5. Cell numbers quantified for control and Cep55^−/−, respectively, were 2585 and 1238 (E13.5), and 2136 and 1238 (E16.5). k Immunohistochemistry of Cep55^+/+ and Cep55^−/− mouse cortices for the basal neural progenitor marker Tbr2. l Left: quantification of the percentage of Tbr2-positive (Tbr2+) nuclei in cortical sections like those in k. n = 3 mice per genotype and per stage; cell numbers quantified for control and Cep55^−/−, respectively, were 1840 and 1080 (E13.5), and 3010 and 1835 (E16.5). Right: Distribution of Tbr2+ nuclei in the VZ and subventricular zone (SVZ). P-values calculated using Student’s two-tailed unpaired t-test. All bar charts show mean ± SD. Source data for c, g, h, j, l are provided as a Source Data file.

Fig. 4. Cep55 deletion results in binucleated cells in the cortex.

a Scheme showing preparation of cells from E16.5 cortices for flow cytometry analysis. b Dot plots showing Ki67 staining and DNA content in Cep55^+/+ and Cep55^−/− cortical cells. Histograms show DNA content of the Ki67-negative (Ki67−) cell populations in the lower boxes. c Quantifications from b. n = 3 and 6 mice per genotype, respectively; n cells > 3000. CTR, control; ^−/−, Cep55^−/−. P-values calculated using Student’s two-tailed unpaired t-test. d, e Confocal optical section of E16.5 cortex from Cep55^−/− brain stained with DAPI (DNA), plasma membrane (PM) stain, and active caspase 3 (C3A). Left boxed areas in d and e are magnified on the right. PM-3D rendering highlights connection of the two cells by a channel. f Top, enlargements of boxed areas from PM and C3A panels in e. Below, corresponding plots showing the fluorescence intensity along the dotted lines, in arbitrary units. Note that the PM appears interrupted. g Left: Immunohistochemistry of P0 Cep55^+/+ and Cep55^−/− cortices for the neuronal marker NeuN. Magenta arrows, NeuN-positive doublets; black arrow, abnormally large nucleus. Right: Quantification of abnormal NeuN-positive (NeuN+) nuclei shown in g. Control includes Cep55^+/+ and Cep55^+/− mice. n = 3 mice per genotype; n = 325 and 284 cells quantified, respectively. P-values calculated using Student’s two-tailed unpaired t-test. h Confocal 3D images of E13.5 cortex from brains of the indicated genotypes stained with DAPI (DNA), Aurora B (green), and MKLP1 (Magenta). Boxed areas are magnified at right. i Quantification of intercellular bridges (ICBs) as shown in boxed areas in h. n = 3 mice per genotype; n = 95 and 59 ICBs, respectively. P-values calculated using Student’s two-tailed unpaired t-test. All bar charts show mean ± SD. Source data for c, g, and i are provided as a Source Data file.

Fig. 5. Cep55 is required for abscission in neural progenitors but not in primary fibroblasts.

a Timeline for Cep55 deletion in Cep55^F/F; R26-Cre^ERT2/+ neural progenitor cells (NPCs), using 4-hydroxytamoxifen (4OHT) to induce recombination or ethanol (EtOH) as vehicle control. IFM, immunofluorescence microscopy; PCR, polymerase chain reaction; qPCR, quantitative PCR; WB, western blotting. b NPCs treated as shown in a were stained for DNA (DAPI) and α-tubulin. Arrows and arrowheads indicate binucleated and pyknotic cells, respectively, quantified in c. n = 3 mice per condition; 1000 cells quantified per condition. *P-values were calculated by one-way ANOVA followed by Tukey’s multiple comparisons test as follows: binucleated cells: P = 0.9703 for FF vs. FF + EtOH, P < 0.0001 for FF vs. FF + 4OHT, P < 0.0001 for FF + EtOH vs. FF + 4OHT; pyknotic cells: P = 0.9673 for FF vs. FF + EtOH, P = 0.0126 for FF vs. FF + 4OHT, P = 0.0165 for FF + EtOH vs. FF + 4OHT. d Timeline for culture and analysis of mouse primary tail tip fibroblasts (TTFs). FACS, fluorescence-activated cell sorting. e TTFs of the indicated genotypes collected as in d were stained as in b. Arrows indicate binucleated cells, quantified in f. n = 3 and 4 mice, respectively; 1300 Cep55^+/+ cells and 1800 Cep55^−/− cells quantified. g Representative time-lapse images of NPCs treated as in a. Arrowhead in the upper panel indicates the intercellular bridge; arrowhead in the lower panel indicates the attempt to divide possibly by cytofission. h Quantification of dividing NPCs as defined in g. n = 3 mice per condition; 56 control cells (Cep55^F/F + EtOH) and 32 recombined cells (Cep55^F/F + 4OHT) were quantified. i Representative time-lapse images of TTFs cultured as in d. Arrowheads indicate the intercellular bridge. j Quantification of dividing Cep55^+/+ and Cep55^−/− TTFs as defined in i. n = 3 mice per genotype; 153 and 171 cells quantified, respectively. All bar charts show mean ± SD. P-values calculated using Student’s two-tailed unpaired t-test in f, h, j. Cep55^F/F indicates Cep55^F/F; R26-Cre^ERT2/+ allele in b, c, g, h, Cep55^−/− indicates Cep55^tm1a/tm1a allele in e and f, and Cep55^tm1d/tm1d allele in i and j. Source data for c, f, h, j are provided as a Source Data file.

Fig. 6. ESCRTs are recruited at the MB of neural progenitors but are absent in Cep55-knockout fibroblasts.

a Quantification (mean ± SD) of MBs with Cep55 present or absent in the indicated cell types. n = 3 mice per cell type; 193, 237, and 214 cells quantified, respectively. b–d Immunofluorescence images of intercellular bridges in the indicated cell types stained for DNA (DAPI), α-tubulin, MKLP1, and Cep55. Boxed areas are magnified on the right. Scale bar in magnified regions, 2 µm. Punctate cytoplasmic signal is nonspecific staining. e Plots showing the fluorescence intensity along the intercellular bridge, in arbitrary units (AU), from TTFs as in c and d. Ten cells per genotype quantified. SEM are shown. f Quantification (mean ± SD) of MBs with Chmp2B present or absent in the indicated cell types. n = 3 mice per cell type; 90, 151, 133, 86, and 67 cells quantified. P-values for TTFs were calculated using two-way ANOVA, each mean compared with control Cep55^+/+ (plastic), followed by Dunnett’s multiple comparisons test. For “Chmp2B present” category, P-values are: 0.2057, 0.0001, and 0.0004, respectively. g–k Immunofluorescence images of the indicated cell types undergoing abscission stained for DNA (DAPI), α-tubulin, MKLP1, and Chmp2B. Boxed areas are magnified below (g–i) or above (j, k). Scale bar, 2 µm. l Quantification of Chmp2B signal (mean ± SD) in TTFs grown on glass or poly-l-lysine (PLL)-coated glass coverslips as in h–k. For Cep55^+/+ TTFs, only Chmp2B-positive cells were analyzed. n = 8, 9, 10, and 12 cells analyzed. P-value calculated using Student’s two-tailed unpaired t-test. m Time-lapse images of control and Cep55^−/− TTFs expressing mChmp4B-EGFP and stained with Sir-tubulin to visualize the microtubules of the ICB. The white arrowheads indicate the site of abscission on the intercellular bridge. n Quantification of mChmp4B-EGFP at the intercellular bridge as in m. Individual pairwise comparisons were performed using χ²-test (one-sided). The omnibus χ²-test P-value is < 0.0001. n = 25, 26, 23, and 19 cells were imaged. Source data for a, e, f, l, and n are provided as a Source Data file.

Fig. 7. ESCRTs are required for abscission in neural progenitors but not in primary fibroblasts.

a Scheme of Chmp4B knockdown in Chmp4C^−/− fibroblasts and wild-type NPCs. siRNA, small interfering RNA. b Chmp4C^−/− fibroblasts were transfected with the indicated siRNAs for 60 h as in a and extracts analyzed for Chmp4B and α-tubulin or actin. CTR, control; si, small interfering RNA; n = 1 and 2 independent experiments, respectively. c Time-lapse images of Chmp4C^−/− fibroblasts depleted of Chmp4B as shown in a. Arrowheads indicate the intercellular bridge. d Quantification of abscission success or failure in dividing fibroblasts, as defined in c. n = 3 mice per condition; 117, 127, 128, 125, 163, and 115 cells quantified, respectively. P-values were calculated using two-way ANOVA, each mean compared with Chmp4C^+/+ on plastic, followed by Dunnett’s multiple comparisons test. e Time-lapse images of wild-type NPCs depleted of Chmp4B as shown in a. f Quantification of abscission success or failure in dividing NPCs, as defined in e. n = 3 mice per condition; 62 and 45 cells quantified respectively. P-values calculated using Student’s two-tailed unpaired t-test. g Model of the different abscission requirements in vivo. Red crosses indicate dying cells. All bar charts show mean ± SD. Source data for b, d, and f are provided as a Source Data file.