Fzr/Cdh1 Promotes the Differentiation of Neural Stem Cell Lineages in Drosophila

How stem cells and progenitors balance between self-renewal and differentiation is a central issue of stem cell biology. Here, we describe a novel and essential function of Drosophila Fzr/Cdh1, an evolutionary conserved protein, during the differentiation of neural stem cell (NSC) lineages in the central nervous system. We show that Fzr, a known co-activator of Anaphase Promoting Complex/Cyclosome (APC/C) ubiquitin ligase, promotes the production of neurons from neural progenitors called ganglion mother cells (GMCs). However, knockdown of APC/C subunit Ida or another APC/C co-activator CDC20 does not similarly impair GMC-neuron transition. We also observe a concomitant loss of differentiation factor Prospero expression and ectopic accumulation of mitotic kinase Polo in fzr mutant clones, strongly supporting the impairment of GMC to neuron differentiation. Besides functioning in GMCs, Fzr is also present in NSCs to facilitate the production of intermediate neural progenitors from NSCs. Taken together, Fzr plays a novel function in promoting differentiation programs during Drosophila NSC lineage development. Given that human Fzr is inactivated in multiple types of human cancers including brain tumors and that Fzr regulates neurotoxicity in various models of neurodegenerative diseases, our study on the role of Fzr in turning off proliferation in neuronal cells may provide insights into how Fzr deficits may contribute to human neurodegenerative diseases and tumors.

How stem cells and progenitors balance between self-renewal and differentiation is a central issue of stem cell biology. Here, we describe a novel and essential function of Drosophila Fzr/Cdh1, an evolutionary conserved protein, during the differentiation of neural stem cell (NSC) lineages in the central nervous system. We show that Fzr, a known co-activator of Anaphase Promoting Complex/Cyclosome (APC/C) ubiquitin ligase, promotes the production of neurons from neural progenitors called ganglion mother cells (GMCs). However, knockdown of APC/C subunit Ida or another APC/C co-activator CDC20 does not similarly impair GMC-neuron transition. We also observe a concomitant loss of differentiation factor Prospero expression and ectopic accumulation of mitotic kinase Polo in fzr mutant clones, strongly supporting the impairment of GMC to neuron differentiation. Besides functioning in GMCs, Fzr is also present in NSCs to facilitate the production of intermediate neural progenitors from NSCs. Taken together, Fzr plays a novel function in promoting differentiation programs during Drosophila NSC lineage development. Given that human Fzr is inactivated in multiple types of human cancers including brain tumors and that Fzr regulates neurotoxicity in various models of neurodegenerative diseases, our study on the role of Fzr in turning off proliferation in neuronal cells may provide insights into how Fzr deficits may contribute to human neurodegenerative diseases and tumors.

INTRODUCTION
Understanding how stem cells maintain their self-renewal capacity and how their progeny differentiate into specific fates are essential to comprehend developmental processes as well as to exploit the therapeutic potential of stem cells for regenerative medicines and cancer treatments. The Drosophila larval brain NSCs, or neuroblasts, have emerged as a fertile model for studying stem cell self-renewal and differentiation in vivo . In the Drosophila central nervous system, type I NSCs divide asymmetrically to self-renew and to generate a smaller daughter cell called the GMC that only divides once to give rise to two terminally differentiated neurons or glial cells (reviewed in Homem et al., 2015). Besides type I NSCs, 8 type II NSCs located bilaterally in the central brain divide asymmetrically to self-renew and generate intermediate neural progenitors (INPs), which can divide 4-6 times to give rise to GMCs and neurons (Bello et al., 2008;Boone and Doe, 2008;Bowman et al., 2008). Newly generated INPs need to differentiate into mature INPs before division. The transcription factor Earmuff (Erm), a homolog of the vertebrate Forebrain embryonic zinc-finger family, is required for INP maturation (Bowman et al., 2008;Weng et al., 2010). Erm works together with the SWI/SNF chromatin-remodeling complex Brahma (Brm) in immature INPs to restrain the developmental potential of INPs (Eroglu et al., 2014;Koe et al., 2014;Janssens et al., 2017;Liu et al., 2017). Several studies in both mammals and Drosophila have indicated that neuronal differentiation is actively maintained. To date, only a few factors have been implicated in maintaining Drosophila neuronal differentiation. These factors include the homeodomain transcription factor Prospero (Pros, human homolog Prox1) (Caussinus and Gonzalez, 2005;Choksi et al., 2006;Maurange et al., 2008), RNA splicing factor Midlife Crisis (Mlc) (Carney et al., 2013), BTB-zinc finger transcriptional factor longitudinals lacking (Lola) (Southall et al., 2014), and the zinc finger transcription factor Nervous fingers 1 (Nerfin-1) (Froldi et al., 2015). In the absence of Pros, GMCs revert to NSCs instead of committing to differentiation (Betschinger et al., 2006;Choksi et al., 2006;Lee et al., 2006). Mlc regulates expression and splicing of Pros and the loss of Mlc leads to the accumulation of ectopic NSC-like cells originating from dedifferentiated neurons, but these NSC-like cells are stalled during cell cycle and do not form tumors (Carney et al., 2013). Lola, on the other hand, acts redundantly with Pros. The loss of Lola is associated with neuron-to-NSC reversion and tumorigenesis in the optic lobes, but not in the central brain due to the compensatory functions of Pros (Southall et al., 2014). Recently, Nerfin-1 has been shown to function downstream of Pros to maintain neuronal differentiation. Neurons in nerfin-1 mutants first increase their cellular size, then switch off neuronal program and start to express the NSC-identity program (Froldi et al., 2015). However, it remains elusive whether other cellular factors play a role in differentiation of NSC lineages.
Here, we describe a novel and essential role for Drosophila Fzr in GMC-to-neuron transition in both type I and type II NSC lineages of the developing larval brains. In fzr − mutant clones, GMC population expands at the expense of neurons. We also observe a concomitant loss of Pros expression and ectopic accumulation of Polo in mutant clones, suggesting the impairment of GMC to neuron differentiation. The localization of EGFP-Fzr BAC in late GMCs is consistent with its novel function in GMCs. Moreover, Fzr also regulates NSC to INP transition, through genetic interaction with Brm and Erm.
Experiments with mutants were performed at 25 • C, and experiments for RNAi-mediated knockdown or overexpression were carried out at 29 • C.

Clonal Analysis
MARCM clones were generated as previously described (Lee et al., 1999). Briefly, the larvae were heat-shocked twice at 37 • C for 2 hours (h) each, shortly after larval hatching (ALH) and at 10-16 h after the first heat shock. Larvae were further aged for another 3 days at 25 • C before dissection.

Immunochemistry
Wandering third instar larval brains were dissected in cold PBS and fixed in 4% EM-grade formaldehyde in PBS at room temperature (RT) for 22 minutes (min), following by three washes in 0.3% Triton-X in PBS (PBST). The sample was incubated with blocking buffer (3% BSA in PBST) for 45 min at RT, followed by incubation with primary antibody mixture overnight at 4 • C. After three washes in PBST, larval brains were incubated with secondary antibody mixture for 90 min at RT, followed by two washes, and mounted in a glycerol based mounting medium (10 mg/ml of p-Phenylenediamine in PBS, 1:10 dilution with glycerol). For staining of DNA, additional incubation of the sample with diluted Topro-3 in PBST was performed just before adding mounting medium. Samples were imaged with Zeiss LSM 710 confocal microscopy and images were processed with Zeiss black software.

Generation of Transgenic Flies
BAC line (CH321-61L05) was used to generate EGFP-Fzr BAC genomic construct according to the method described previously (Zhang et al., 2019). Briefly, a fragment containing enhanced GFP (EGFP) tag and Kanamycin-resistant gene was amplified from PL-452 EGFP vector and inserted to the N-terminus of fzr coding sequence by recombination using SW102 electrocompetent cells. The resulting cassette was then electroporated into ara-inducible Cre carrying SW106 electrocompetent bacteria to remove the Kanamycin resistant gene. The correct clone carrying EGFP-Fzr BAC was electroporated into EPI300 cells for amplification and the BAC DNA was purified using BACMAX kit (Epicentre Biotechnologies Cat# BMAX044). The EGFP-Fzr BAC was sent to BestGene, Inc. for injection into y 1 w 67 c 23 ; P{CaryP}attP40 (estimated cytosite 25C6) background. The primers used for recombination are listed in Supplementary Table 1.

Statistical Analysis
GraphPad Prism 6 software was used for statistical analysis. All data were presented as mean ± SD. Unpaired two-tail t-tests were used for two sample comparisons and one-way ANOVA for comparison of more than two groups. In ANOVA, Dunnett's post hoc test was used to obtain the P-values for pairwise comparison. In this work, comparisons were performed against wild-type or control, unless otherwise indicated by a line between two genotypes. A value of P ≤ 0.05 was considered as statistically significant, * indicated P ≤ 0.05, * * indicates P ≤ 0.01, * * * indicates P ≤ 0.001, and * * * * indicates P ≤ 0.0001. P > 0.05 was considered as statistically non-significant (ns).

Loss of Fzr in NSC Lineages Results in Ectopic Progenitor GMCs
In order to evaluate the possible function of Fzr in NSC lineage development, we generated mosaic analysis with a repressible cell marker (MARCM) clones for two known mutant alleles of fzr, named fzr A and fzr B , which were isolated by ethyl methanesulfonate (EMS) mutagenesis on X-chromosome (Yamamoto et al., 2014). Both alleles failed to complement fzr G0326 , the hypomorphic allele with a P{lacZ} element inserted in the first intron of fzr, and were rescued by Dp(1;3)DC120 duplication regarding lethality (this study and Yamamoto et al., 2014). We have tried but failed to characterize the precise lesion of fzr A and fzr B . Whereas all the control type I NSC lineages (n = 40) contain a single Dpn + Ase + NSC and 4-5 Dpn − Ase + GMCs that undergo terminal division, fzr A and fzr B (collectively called fzr − ) mutant clones contain a single NSC but a large number of ectopic GMCs (17.8 ± 6.38 and 19.6 ± 8.20, respectively) ( Figures 1A,B). The ectopic GMC phenotype observed in fzr − mutants were fully rescued by expressing an UAS-Fzr-HA transgene driven by tub-Gal4 from MARCM driver as well as by the insertion of Dp(1;3)DC120, a genomic fragment containing the fzr locus (Figures 1A,B).
Next, we assessed whether ectopic GMCs were also observed in type II NSC lineages in fzr − mutants. There are 8 type II NSC lineages in each larval brain lobe. The defining characteristic of type II NSC clones is that they produce INPs, which can divide 4-6 times to generate GMCs. Within type II lineages, the NSCs can be recognized by their large size (10-14 µm in diameter) and positive for Dpn but negative for Ase. INPs can be identified by their small size (3-4 µm in diameter) and are either Dpn + Ase + (mature INPs) or Dpn − Ase + and Dpn − Ase − (immature INPs). We further analyzed the number of immature FIGURE 1 | Ectopic GMCs are generated upon Fzr deletion. (A) Type I NSC clones of MARCM driver control (FRT19A), fzr A , fzr B , fzr 8F3 , fzr A + Dp(1;3)DC120, fzr B + Dp(1;3)DC120, fzr 8F3 + Dp(1;3)DC120, fzr A + tub > Fzr-HA, fzr B + tub > Fzr-HA, and fzr 8F3 + tub > Fzr-HA were labeled with Dpn, Ase and CD8. tub > Fzr-HA refers to UAS-Fzr-HA driven by tub-Gal4 from MARCM driver, so Fzr-HA is expressed only in MARCM clones that lose tub-Gal80 upon mitotic recombination. (B,C) Quantifications of GMC number (B) per type I MARCM clones and the percentage of MARCM clones with two or more NSCs (C) for (A).  Figures 1A,B). Since we did not detect any change in the number of mature and immature INPs (Supplementary Figures 1C,E-G), ectopic Dpn − Ase + cells observed in type II fzr − clones are most likely GMCs. Together, these results suggest that Fzr regulates the homeostasis of GMCs in both type I and type II NSC lineages. Consistently, the knockdown of Fzr in NSCs (and their immediate progenies due to perdurance of Gal4 in progenies) using RNA interference (RNAi) with pan-NSC driver (wor-Gal4;UAS-Dcr2) resulted in ectopic GMCs (Figures 1E,F) that are often undergoing cell division (Figures 1E,G). Similarly, the knockdown of Fzr specifically in type I or type II NSCs (and immediate progenies of NSCs due to Gal4 protein perdurance) also leads to an increased number of GMCs (Supplementary Figures 1H-K).
Moreover, we also observed ectopic GMCs phenotype in clones from another fzr − mutant, fzr 8F3 as well as hemizygous mutant fzr G0418 /Y (Figures 1A-D). The EMS fzr 8F3 mutant contains a nonsense mutation (Trp214 > Stop) for fzr and likely produces short unstable peptides (Silies and Klämbt, 2010). The fzr G0418 has the P{lacW} element inserted in the 5 end of fzr gene, and is hypomorphic allele of fzr (Jacobs et al., 2002). Besides having ectopic GMCs, a few fzr B clones also contained ectopic small Dpn + Ase + NSClike cells (Figure 1C). Similarly, in fzr 8F3 mutant clones, 45.8% (n = 24) and 80% (n = 20) of mutant type I and type II clones also contained ectopic NSCs (Figures 1A,C  and Supplementary Figures 1A,D). While the ectopic GMC phenotype observed in fzr 8F3 clone was completely rescued by the insertion of genomic fragment Dp(1;3)DC120, the ectopic NSC phenotype was only partially rescued (Figures 1A,C  and Supplementary Figures 1A,D), suggesting an additional background mutation present in the fzr 8F3 mutant might partially contribute to the ectopic NSC phenotype. Hereafter we focus on the function of Fzr in regulating GMC population in type I NSC clones.
In conclusion, Fzr plays an important role in NSC lineage progression in developing larval brains and that the Fzr loss results in the formation of ectopic GMCs.

Fzr − Neural Lineages Do Not Display Any Disturbance of NSC Asymmetric Division or Neuronal Dedifferentiation
To further characterize the cell fate of GMCs accumulated upon the loss of Fzr, we examined the expression of Nerfin-1, which is expressed in late GMCs and newly generated neurons. Whereas each control type I NSC clone contains only on average 1.09 ± 0.702 Ase + Nerfin-1 + late GMCs and around 12.2 ± 4.82 Ase − Nerfin-1 + early neurons, there were 25.7 ± 8.52 and 24.0 ± 7.31 Ase + Nerfin-1 + late GMCs and 2.26 ± 2.47 and 2.65 ± 3.42 Ase − Nerfin-1 + early neurons in fzr A and fzr B mutant clones, respectively (Figures 2A-C). Wild-type neurons express Embryonic lethal abnormal vision (Elav) but not Asense ( Figure 2D). Strikingly, we observed a population of Elav + Ase + cells in fzr − mutants with a concomitant reduction of Elav + Ase − neurons (Figures 2D-F). Consistently, upon the loss of Fzr in type I MARCM clones, some of the Elav + Ase + cells are undergoing mitosis, as judged by the presence of phosphor-Histone 3 (PH3) (Figures 2D,G). Altogether, these results indicate that loss of Fzr in NSC lineages results in the accumulation of dividing GMCs at the expense of generating neurons undergoing differentiation.
There are three non-mutually exclusive possibilities that could contribute to this interesting phenotype: (1) an ectopic production of GMCs from disrupted asymmetric divisions or extended NSC division; and/or (2) dedifferentiation of neurons to GMCs; and/or (3) failures during GMC differentiation that lead to their ectopic cell divisions.
To address the first possibility that the disturbance in NSC division results in production of excess GMCs, we examined the localization of polarity proteins that are essential for NSC division. We observed that the cellular localization of apically localized aPKC and basally localized Miranda (Mira) in fzr 8F3 dividing NSCs are intact and similar to that of the control NSCs ( Figure 2H). Similarly, Mira in both control as well as fzr A and fzr B NSCs are correctly localized (Supplementary Figure 2A). The localization of another apically localized protein Bazooka (Baz, Drosophila homolog of mammalian Par3) are also intact in fzr 8F3 mutant NSCs (Figure 2I). In addition, the localization of mitotic kinase Polo in dividing NSCs seems unaffected in fzr A and fzr B NSCs (Supplementary Figure 2B). Besides, the number of progeny cells from a single NSC remains similar upon Fzr loss (Supplementary Figure 1L) and that no elevated cell-death, as judged by the staining of cleaved caspase 3, was observed in fzr A clones (Supplementary Figure 2C). Together, these results suggest that NSC division seems normal upon loss of Fzr and that disturbances in NSC division are unlikely the cause of GMC accumulation.
To test whether neuronal dedifferentiation accounts for the ectopic GMCs, we knocked down Fzr in different populations of NSCs lineages with fzr RNAi with different Gal4 drivers. While the loss of Fzr in whole NSC lineage using elav-Gal4 or pros-Gal4 drivers resulted in ectopic GMCs (Figures 2J,K and Supplementary Figure 2D), the loss of Fzr in late GMCs and early neurons by nerfin-1-Gal4 driver did not result in ectopic GMCs within nerfin-1 expressing clones (labeled by GFP) (Figures 2L,M). Moreover, using an enhancer trap in which betagalactosidase (β-Gal) is inserted at fzr locus (fzr G0418 ), we did not detect the expression of Fzr in Elav + neurons (Supplementary Figure 2E), suggesting that Fzr does not function in neurons to maintain neuronal differentiation. In summary, our results indicate that neuronal dedifferentiation is unlikely the cause of GMC accumulation.
Interestingly, we observed a high expression of Fzr in glial cells (Supplementary Figure 2F). However, upon the knockdown of Fzr in glial cells by fzr RNAi with glial driver (repo-Gal4), GMC number remains unchanged (Supplementary Figures 2G-I), indicating that GMC number is not regulated by Fzr in glial cells.
In summary, our results suggest that ectopic GMC accumulation upon the loss of Fzr is likely caused by the failed differentiation of GMCs that leads to ectopic GMC divisions.

Fzr − Neural Lineages Display Defects in GMC-to-Neuron Transition
To test if there are any defects during the transition of GMCs to neurons, we assessed the expression of differentiation factor Prospero (Pros) (Chu-Lagraff et al., 1991;Choksi et al., 2006) in fzr − NSC clones. In most fzr − clones, the expression of Pros is substantially reduced in comparison to those of control clones ( Figure 3A). Moreover, the expression of Polo kinase, which promotes cell division and stemness (Sunkel and Glover, 1988;Glover, 2005), remained high in ectopic GMCs of fzr − clones ( Figure 3B). All these results suggest that the ectopic mutant GMCs maintain their undifferentiated states and are unable to differentiate into neurons.
In addition, Fzr displayed moderate staining in wild-type NSCs and weak staining in GMCs (Figure 3C, asterisks and yellow arrows, respectively), consistent with our model that Fzr functions within GMCs to promote the transition of GMCs into neurons. To further characterize the expression pattern of endogenous Fzr in the nervous system, we attempted to generate anti-Fzr antibodies for immunostaining without success (data not shown). Next, we turned to labeling Fzr protein with enhanced GFP (EGFP) within its endogenous locus, hereafter called EGFP-Fzr BAC . Consistent with the localization of Fzr shown by the enhancer trap, EGFP-Fzr BAC displayed strong staining in glial cells, as judged by their glial morphology and cortical location ( Figure 3D, white arrows) as well as being positive for the glial marker Repo (Supplementary Figure 3A). Besides, EGFP-Fzr BAC also displayed strong localization in late GMCs ( Figure 3E, yellow arrows), minimal level in early GMCs ( Figure 3E, blue arrows, as judged by their small size, being Ase + Mira weak cells, and their immediate proximity to the big Ase + Mira + type I NSCs) and weak level in NSCs (Figure 3E, white asterisks).

Fzr Loss Enhances Ectopic Type II NSCs Resulted From Downregulation of Differentiation Factors Erm or Brm
Our data shows that the loss of Fzr also results in the formation of ectopic NSCs in addition to excess GMCs (Figures 1A-C and Supplementary Figures 1A-D). Brahma (Brm) and Earmuff (Erm) are important to suppress dedifferentiation of INPs, and the loss of either Brm or Erm results in the reversion of INP to type II NSCs. Next we wondered if Fzr genetically interacted with Brm or Erm to suppress INP dedifferentiation. While Fzr knockdown alone using NSC-specific driver did not cause ectopic type II NSCs (8 type II NSCs per brain lobe similar to a control brain lobe, Figures 4A,B), the loss of Brm resulted in ectopic type II NSCs (15.2 ± 1.92 NSC, n = 5). However, knockdown of Fzr dramatically enhanced dedifferentiation defects associated with Brm loss to 81.8 ± 12.8 NSCs per brain lobe (Figures 4A,B). Similarly, the knockdown of Fzr using NSC-specific driver (insc-Gal4) substantially enhanced the ectopic type II NSCs of Erm knockdown (Figures 4C,D, 78.7 ± 7.30 NSCs, n = 11 BLs in erm RNAi + control RNAi vs. 311.9 ± 169.6 NSCs, n = 20 BLs in erm RNAi + fzr RNAi ). The overexpression of Rca1, negative regulator of Fzr-dependent functions of APC/C complex, as well as Fzr overexpression did not affect the severity of Erm loss, suggesting Fzr might function independent of APC/C complex in the regulation of INP-NSC homeostasis (Figures 4C,D). Interestingly, the knockdown of Fzr using INP driver (erm-Gal4, UAS-CD8-GFP) resulted in ectopic type II NSCs from 9.5% of INP clones (n = 105), while none of the control INP clones contained any type II NSCs ( Figure 4E). Due to the unavailability of independent RNAi lines targeting different fzr sequences and the lethality of homozygous fzr mutants in the early embryonic or larval stages, we were unable to confirm genetic relationship of Fzr and Brm or Erm in NSC-INP transition with additional RNAi lines or mutants. Moreover, knockdown of the potential off-target of fzr RNAi (GD#25553), bicoid (bcd), by RNAi under the control of the same NSC-specific driver (insc-Gal4) did not lead to as strong enhancement of INP dedifferentiation associated with Erm loss, in comparison to the effect of fzr RNAi (Figures 4C,D), suggesting that loss of fzr partially contributed to the genetic enhancement observed in the double knockdown fzr RNAi and erm RNAi .
Altogether, these results suggest that beside its roles in regulating GMC-to-neuron transition, Fzr genetically interacts with Brm and Erm to promote NSC-INP transition.

DISCUSSION
Here, we show that Fzr/Cdh1, an established co-activator of APC/C ubiquitin ligase known for its roles in regulating cell cycle or post-mitotic functions in terminally differentiated neurons or glial cells, promotes the commitment of neural progenitor GMCs to the production of terminally differentiated neurons (Figures 5A,B). In this study, we present multiple lines of evidence to support the function of Fzr in regulating GMC-toneuron differentiation: (1) the presence of Fzr in late GMCs; (2) the accumulation of GMCs at the expense of neurons in fzr − mutant NSC clones; and (3) the loss of differentiation factor Pros and ectopic accumulation of mitotic Polo kinase in NSC progenies upon loss of fzr. Beside functioning in GMCs, Fzr is also present in low level in NSCs to regulate NSC-to-INP transition of type II NSC lineages (Figures 5A,C).
In this study, we observed a strong expression of EGFP-Fzr in the optic lobe ( Figure 3D). However, the function of Fzr in optic lobes is currently unclear. Recently, the APC/C Fzr complex has been reported to promote retinal differentiation in Drosophila eye imaginal discs and thus the formation of adult eyes (Martins et al., 2017). In eye discs, Fzr modulates Wingless Signaling via Nek2 degradation, which is essential for progenitor cells to differentiate. As such, it would be interesting to examine if Fzr might regulate GMC differentiation in NSC lineages of the central brain through a similar mechanism.
In this study, we provide putative evidence that Fzr interacts with Erm and Brm to regulate NSC-INP balance. Interestingly, Ets2, the mammalian homolog of the master regulator of type II lineages PntP1, is stabilized upon Cdh1 deficiency and is proposed to be substrate of APC/C Cdh1 complex (Li et al., 2008). However, it remains to be determined how Fzr, Erm, or Brm cooperate to regulate NSC-INP transition in type II NSC lineages and if the underlining mechanisms for Fzr's function in GMCneuron and NSC-INP transitions are shared.
In mammals, Cdh1/Fzr and several core components of APC/C complex are highly expressed in post-mitotic neurons (Gieffers et al., 1999) and Cdh1/Fzr participates in the regulation of neuronal axonal and dendritic growth, synapses, metabolism and survival of neurons (Eguren et al., 2011). Besides, Fzr/Cdh1 functions to prevent replicative stress and p53-dependent cell death in neural progenitors (Eguren et al., 2013). In this study, we uncover a new role of Fzr in Drosophila nervous system: ensuring the commitment of neural progenitors to differentiation cascade. Large scale human cancer tissue arrays and prognostic analyses have indicated that loss of APC Cdh1 function correlates with various human carcinogenesis, including brain tumors (reviewed in Qiao et al., 2010). Besides, inactivation of Cdh1 has been implicated in excitation-mediated neuronal cell death in neurological disorders such as Alzheimer disease (Maestre et al., 2008). Our study on the role of Fzr in turning off proliferation in neuronal cells may provide insight into how Fzr deficit may contribute to human neurodegenerative diseases and tumors.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material (see Supplementary Data Sheet S2 for the numerical data used in this study).

AUTHOR CONTRIBUTIONS
PL: conceptualization, data curation, formal analysis, investigation, methodology, writing -the original draft, and writing -review and editing. HW: conceptualization, funding acquisition, project administration, supervision, writing -review and editing the original draft, and writing -review and editing.

FUNDING
This work was supported by Duke-NUS Signature Research Program funded by Ministry of Health and Ministry of Education tier 2 (MOE2017-T2-2-025) to HW. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.