Robust elimination of genome-damaged cells safeguards against brain somatic aneuploidy following Knl1 deletion

Abstract

The brain is a genomic mosaic shaped by cellular responses to genome damage. Here, we manipulate somatic genome stability by conditional Knl1 deletion from embryonic mouse brain. KNL1 mutations cause microcephaly and KNL1 mediates the spindle assembly checkpoint, a safeguard against chromosome missegregation and aneuploidy. We find that following Knl1 deletion, segregation errors in mitotic neural progenitor cells give rise to DNA damage on the missegregated chromosomes. This triggers rapid p53 activation and robust apoptotic and microglial phagocytic responses that extensively eliminate cells with somatic genome damage, thus causing microcephaly. By leaving only karyotypically normal progenitors to continue dividing, these mechanisms provide a second safeguard against brain somatic aneuploidy. Without Knl1 or p53-dependent safeguards, genome-damaged cells are not cleared, alleviating microcephaly, but paradoxically leading to total pre-weaning lethality. Thus, mitotic genome damage activates robust responses to eliminate somatic mutant cells, which if left unpurged, can impact brain and organismal fitness.

Fig. 1

Knl1 conditional deletion from cortical NPCs led to microcephaly and NPC loss. a Dorsal view of Knl1^+l+;hGFAP-Cre and Knl1^fl/fl;hGFAP-Cre (cKO) P4 brain. Cortical area was significantly reduced in cKO compared to littermate control (ctrl) (mean, two-tailed unpaired t-test, ctrl: n = 4, cKO: n = 3 animals, scale bar: 1 mm). b DAPI staining in coronal section of Knl1^+l+;hGFAP-Cre and Knl1^fl/fl;hGFAP-Cre (cKO) P4 brain. Cortical plate (CP) thickness was significantly reduced in cKO compared to ctrl (mean, two-tailed unpaired t-test, ctrl: n = 7, cKO: n = 4 animals, scale bar: 500 µm). c Postnatal survival analysis showed reduced survival in cKO (n = 11) compared to ctrl (n = 25 animals). d Cortical layer marker analysis at P4 revealed no significant change in BCL11B+ (L5, magenta) or TBR1+ (L6, red) neurons but a significant reduction in number of CUX1+ (L2–4, green) and POU3F2+ (L2–5, cyan) neurons in cKO compared to ctrl (mean, two-tailed unpaired t-test, TBR1, ctrl: n = 3, cKO: n = 4, BCL11B, ctrl: n = 6, cKO: n = 4, POU3F2, ctrl: n = 3, cKO: n = 3, CUX1, ctrl: n = 6, cKO: n = 4 animals, scale bar: 100 µm). e NPC marker analysis of E13.5 cortex revealed no significant change in cKO. At E15.5 and E16.5, SOX2+ apical progenitors (APs, green) and EOMES+ intermediate progenitors (IPs, magenta) were significantly reduced in number in cKO (mean, two-tailed unpaired t-test, EOMES E13.5, ctrl: n = 3 cKO: n = 5, E15.5, ctrl: n = 4, cKO: n = 4, E16.5, ctrl: n = 3, cKO: n = 3, SOX2 E13.5, ctrl: n = 3, cKO: n = 5, E15.5, ctrl: n = 3, cKO: n = 3, E16.5, ctrl: n = 3, cKO: n = 3 animals, scale bar: 50 µm). Aberrant gaps in immunolabeling (arrowheads) were observed in cKO. MB midbrain, CPu caudate putamen, IZ intermediate zone

Fig. 2

Robust and rapid apoptosis following Knl1 deletion. a Cleaved Caspase 3 (CC3) immunostaining showed extensive apoptosis in E15.5 cKO cortex (scale bar: 100 µm). b DAPI staining revealed an abundance of pyknotic nuclei in E15.5 cKO cortex (scale bar: 50 µm). Many pyknotic nuclei exhibited a stereotyped clustering (arrowheads). c CC3 immunostaining showed that apoptosis was present in E13.5 cKO cortex, one day after the onset of hGFAP-Cre-mediated Knl1 deletion at E12.5. d At E15.5, near the peak of neurogenesis, CC3 staining (green) was extensive in cortical VZ, SVZ, and intermediate zone (IZ), but largely absent from RBFOX3+ (NeuN+) cortical plate (CP) neurons (red) in cKO. e By E17.5, CC3 staining showed only a moderate increase in cKO compared to ctrl (scale bar: 50 µm). f, g Analysis of E16.5 cortex following a 4-h pulse of CldU (red) and a 24-h pulse of EdU (blue, scale bar: 100 µm). CC3 staining (green) showed no apoptosis in CldU+ cells (open arrowheads) in ctrl or cKO, but revealed extensive apoptosis in EdU+ cells (solid arrowheads) in cKO. EdU+ migrating neurons (yellow arrows) were present in the IZ of ctrl but largely absent from cKO (inset scale bar: 20 µm). h Quantification of CC3 colocalization with CldU or EdU (mean, two-tailed unpaired t-test, n = 3 animals). LGE lateral ganglionic eminence, h hours

Fig. 3

DNA damage and activation of p53 and p53 target genes in Knl1 cKO brain. a Unique molecular identifier (UMI) RNA-seq volcano plot comparing cKO (n = 4) to ctrl (n = 7 animals) E15.5 cortex. P-value was calculated with likelihood ratio tests and false discovery rate (FDR) was calculated using the Benjamini–Hochberg procedure. Differentially expressed genes (FDR < 0.05) are indicated in red. b All 19 significantly upregulated genes. p53 target genes (pink), microglial genes (blue), and interneuronal genes (green) are indicated. c Droplet digital RT-PCR (ddRT-PCR) validated upregulation of p53 target genes Ano3, Eda2r, and Pvt1, and downregulation of Eomes in cKO compared to ctrl (mean, two-tailed unpaired t-test, ctrl: n = 7, cKO: n = 4 animals). d Immunostaining of E15.5 cortex revealed significant increases in the number of cells labeled by p53 or DNA damage marker γH2AX in cKO compared to ctrl (mean, two-tailed unpaired t-test, n = 3 animals, scale bar: 100 µm). p53 was localized to nuclei (solid arrowheads) and present in pyknotic cells (open arrowheads, inset scale bar: 5 µm). e Double immunostaining revealed p53 activation in SOX2+ NPCs in E15.5 cKO VZ (arrowheads). The vast majority of cells (89 ± 1.5%) with p53 activation were SOX2+ NPCs (mean, cKO: n = 4 animals). coloc. colocalization

Fig. 4

Missegregated chromosomes with DNA damage in mitotic NPCs following Knl1 deletion. a DAPI staining revealed frequent lagging and bridged chromosomes (open arrowheads), and micronuclei (solid arrowheads) in anaphase NPCs in E15.5 cKO cortex. Missegregated chromosomes and micronuclei were present in the midzone labeled by phospho-Vimentin (pVIM, red, scale bar: 2 µm). b The percentage of anaphase cells with chromosome missegregation was significantly increased in cKO compared to ctrl (mean, two-tailed unpaired t-test, n = 4 animals). c Missegregated chromosomes and micronuclei in cKO were immunopositive for DNA damage marker γH2AX (green). d Dissociated E15.5 cortical NPC culture after 2 days in vitro (DIV) and a 12-h pulse of EdU (red). NPCs cultured from cKO were characterized by DNA bridges (open arrowheads) and micronuclei (solid arrowhead) immunopositive for γH2AX (green). e, f Single cell electrophoresis comet assay of E14.5 cKO and ctrl cortices (scale bar: 10 µm). Comet tail lengths were significantly increased in cKO compared to ctrl (mean, two-tailed unpaired t-test, ctrl: n = 36, cKO: n = 26 cells)

Fig. 5

Co-deletion of Trp53 with Knl1 partially ameliorated Knl1 phenotypes. a Dorsal view of Knl1^fl/+;Trp53^fl/+, Knl1^fl/fl;Trp53^+/+;hGfap-Cre (cKO), and Knl1^fl/fl;Trp53^fl/fl;hGfap-Cre (dKO) P0 brain (scale bar: 1 mm). The significant cortical area loss in cKO compared to littermate control (ctrl) was partially ameliorated in dKO (mean, ANOVA with post-hoc t-test, ctrl: n = 11, cKO: n = 3, dKO: n = 3 animals). b Cortical layer marker analysis at P0 revealed no significant change in BCL11B+ (L5, red) neurons (scale bar: 50 µm). The reduction in CUX1+ (L2–4, green) neuron number in cKO compared to ctrl was partially ameliorated in dKO (mean, ANOVA with post-hoc t-test, ctrl: n = 4, cKO: n = 3, dKO: n = 3 animals). c UMI RNA-seq volcano plot comparing cKO to ctrl, and dKO to ctrl E15.5 cortex (ctrl: n = 7, cKO: n = 4, dKO: n = 4 animals). In the cKO versus ctrl comparison, differentially expressed genes (FDR < 0.05) are indicated in red (left panel). The same genes were labeled in the dKO versus ctrl comparison (right panel). One gene (gene model Gm29260) showed differential expression in dKO (FDR < 0.05) and is indicated in red. The other genes did not show significant differential expression in dKO (FDR ≥ 0.05) and are indicated in black. d Intersectional analysis of p53 target genes identified by p53 ChIP, case studies of p53 transcriptional regulation (Txn. case), genome wide studies of p53 transcriptional regulation (Txn. GW), and ionizing radiation-induced p53 transcriptional regulation (Txn. X-ray). p53-dependent genes induced by ionizing radiation were upregulated in cKO versus ctrl, and showed a significant enrichment (mean, two-tailed paired t-test and hypergeometric test)

Fig. 6

Persistent DNA damage and attenuated apoptosis in Knl1/Trp53 dKO. a In E15.5 dKO brain, DAPI staining revealed lagging chromosomes (open arrowheads) and micronuclei (red solid arrowhead) in anaphase NPCs similar to those in cKO (scale bar: 2 µm). b Single cell electrophoresis comet assay of E14.5 dKO cortices (scale bar: 10 µm). c Comet tail lengths were significantly increased in dKO compared to ctrl (mean, ANOVA with post-hoc t test, ctrl: n = 36, cKO: n = 26, dKO: n = 24 cells). d CC3 immunostaining in E15.5 cortex showed reduced apoptosis in dKO compared to cKO (scale bar: 100 µm). e, f DAPI staining in E15.5 dKO cortex revealed pyknotic nuclei (open arrowheads), which was increased in number in dKO compared to ctrl (scale bar: 10 µm). This increase, however, was significantly smaller than the increase in cKO (mean, ANOVA with post-hoc t-test, ctrl: n = 5, cKO: n = 4, dKO: n = 4 animals). g Immunostaining of E15.5 cortex revealed a significant increase in the number of γH2AX+ cells in dKO compared to cKO or ctrl (mean, ANOVA with post-hoc t-test, ctrl: n = 5, cKO: n = 4, dKO: n = 4 animals), indicating that without p53, apoptosis in dKO was less efficient at clearing genome-damaged cells (scale bar: 100 µm)

Fig. 7

Phagocytosis of apoptotic cells by activated microglial in Knl1 cKO and Knl1/Trp53 dKO. a Developmental analysis by DAPI staining revealed that pyknotic nuclei (open arrowheads) were present at E13.5, peaked in number at E15.5, and eliminated by birth in cKO cortex (mean, two-tailed unpaired t-test, E13.5, ctrl: n = 4 cKO: n = 3, E15.5, ctrl: n = 4, cKO: n = 6, E17.5, n = 3, P0, ctrl: n = 4, cKO: n = 3, P4, n = 4 animals, scale bar: 50 µm). b ADGRE1 (F4/80) immunostaining revealed morphologic activation of microglia in E15.5 cKO and dKO cortex (scale bar: 100 µm). Confocal z-stacking revealed that ADGRE1+ (green) microglia acquired a ballooned morphology. c ADGRE1+ (green) microglia with ballooned morphology phagocytosed dozens of DAPI+ (white) pyknotic nuclei, some of which were CC3+ (red, and open arrowheads) in cKO and dKO cortex at E15.5 (scale bar: 5 µm). d Phagocytic microglia in cKO and dKO were significantly larger in size compared to ctrl microglia. However, the increase in microglial size was significantly attenuated in dKO compared to cKO (mean, ANOVA with post-hoc t-test, n = 3 animals). e The number of AIF1+ (IBA1+) microglia was significantly increased in cKO and dKO compared to ctrl. The increase in microglial number, however, was attenuated in dKO compared to cKO (mean, ANOVA with post-hoc t test, ctrl: n = 5, cKO: n = 4, dKO: n = 3 animals)

Fig. 8

Persistent somatic mutant cells and complete pre-weaning lethality in Knl1/Trp53 dKO. a, b Karyotype analysis of E14.5 NPCs by metaphase spread. Mitotic NPCs in cKO largely had normal karyotypes similar to ctrl. In contrast, 79% of mitotic NPCs in dKO were aneuploid (mean, error bars: ±s.e.m., Fisher’s exact test, ctrl: n = 19, cKO: n = 13, dKO: n = 19 metaphase spreads). c Survival analysis of ctrl, cKO, and dKO showed complete pre-weaning lethality in dKO (ctrl: n = 25, cKO: n = 11, dKO: n = 14 animals). d Schematic summary. In Knl1 cKO, DNA damage on missegregated chromosomes in mitotic NPCs were a potent trigger for rapid p53-dependent apoptotic and downstream microglial mechanisms that extensively cleared the NPCs with segregation errors, leading to massive cell loss and microcephaly. This clearance left only karyotypically normal NPCs to continue dividing. In Knl1/Trp53 dKO, DNA damage on missegregated chromosomes was unable to elicit robust apoptotic or microglial responses, thus leaving genome-damaged and aneuploid NPCs unpurged and able to continue dividing. This alleviated the severity of microcephaly, but paradoxically resulted in complete pre-weaning lethality. Chrom. chromosome