Evolutionarily Conserved Role of Nucleostemin: Controlling Proliferation of Stem/Progenitor Cells during Early Vertebrate Development

Abstract

Nucleostemin (NS) is a putative GTPase expressed preferentially in the nucleoli of neuronal and embryonic stem cells and several cancer cell lines. Transfection and knockdown studies indicated that NS controls the proliferation of these cells by interacting with the p53 tumor suppressor protein and regulating its activity. To assess the physiological role of NS in vivo, we generated a mutant mouse line with a specific gene trap event that inactivates the NS allele. The corresponding NS^−/− embryos died around embryonic day 4. Analyses of NS mutant blastocysts indicated that NS is not required to maintain pluripotency, nucleolar integrity, or survival of the embryonic stem cells. However, the homozygous mutant blastocysts failed to enter S phase even in the absence of functional p53. Haploid insufficiency of NS in mouse embryonic fibroblasts leads to decreased cell proliferation. NS also functions in early amphibian development to control cell proliferation of neural progenitor cells. Our results show that NS has a unique ability, derived from an ancestral function, to control the proliferation rate of stem/progenitor cells in vivo independently of p53.

Stem cells are present throughout embryonic development as well as in several adult organs. They constitute a pool of undifferentiated cells with the remarkable ability to perpetuate through self-renewal while remaining able to terminally differentiate into various mature cell types. Coordinated control of self-renewal and commitment to differentiation is key to maintaining the homeostasis of the stem cell compartment, and its deregulation may contribute to cancer pathogenesis (35). The identification of stem cell-specific proteins and the elucidation of novel regulatory pathways that ensure the integration of these processes are therefore of fundamental importance.

Nucleostemin (NS) was identified because it is highly expressed in rat neuronal stem cells. It is also highly expressed in expanded neurospheres from the adult subventricular zone, in mouse embryonic neuronal stem/progenitor cells between embryonic day 8.5 (E8.5) and E14.5, and in adult bone marrow stem cells (46). Interestingly, NS was markedly down-regulated during cellular differentiation. NS expression declines considerably after E10.5 in the mouse cerebral cortex and is undetectable in lineage-committed B lymphocytes or granulocytes. In culture, NS was found in virtually all rat embryonic cortical stem cells but became undetectable after treatment with ciliary neurotrophic factor, when most cells turn into astrocytes. Importantly, both during neuronal development in vivo and differentiation in vitro, NS protein levels decrease in dividing neuronal cells, whereas proteins that mark cell cycle exit are down-regulated at a later phase. These data indicate that NS down-regulation may lead to cell cycle exit rather than occur as a consequence of cellular differentiation and cell cycle withdrawal (46). Thus, NS is predominantly expressed in stem/progenitor cells and may play an important role in controlling their proliferation. Importantly, NS may also be involved in regulating the proliferation of cancer cells. Indeed, NS was found in a number of human cancer cell lines, such as H1299, U2OS, Soas-2, U937, SW480, 95D, and HEK293, and in malignant renal tissues from patients with clear cell renal cell carcinomas (46, 18, 16). Moreover, NS knockdown experiments in U2OS cancer cells resulted in an increase in noncycling cells (46).

The nucleolus is the subnuclear compartment where rRNA transcription and ribosome assembly occurs. Since NS protein is predominantly found in the nucleoli of undifferentiated cells and cancer cells, its involvement in the regulation of cell proliferation may be indirectly linked to a function in ribosome synthesis. Intuitively, efficient protein synthesis is required to support growth and proliferation. In agreement, links between cell proliferation control and ribosome biogenesis have been demonstrated experimentally (14, 44). However, NS is excluded from the nucleolar domains in which ribosomes are born and appears concentrated in rRNA-free sites within the granular component (33). Alternatively, NS may be associated with other nucleolar functions. Indeed, the notion that this subnuclear structure functions as a mere ribosome factory has recently been challenged. More recent findings showed that the nucleolus also functions as a storehouse for titrating certain proteins and consequently modulating their molecular pathways (30). For instance, mounting evidence highlights a critical role for the nucleolus in the regulation of the p53 pathway, the activity of which is essential for controlling cell proliferation and survival in response to cellular insults (31). In addition to the key role of nucleolar ARF in the activation of p53 in response to oncogenic stress (49), recent studies have shown that deregulation of the expression of a number of nucleolar proteins, such as ribosomal proteins L5, L11, L23, and nucleophosmin (NPM or B23) can affect p53 function (8, 24, 10, 11, 22, 21, 7, 9). An attractive model, in which NS acts to suppress the growth suppressive function of p53 in actively proliferating stem/progenitor cells and cancer cells, has recently been proposed (5). NS may regulate the localization and function of proteins participating in the regulation of p53 activity. Alternatively, NS may control the activity of the p53 pathway via a key role in the maintenance of nucleolar integrity. Indeed, all types of stress that induce p53 appear to disrupt nucleolar integrity, and a model in which impairment of nucleolar structure/function was suggested to stabilize p53 by interfering with its degradation was proposed (36). Finally, regulation of p53 may occur through direct binding because NS was found to interact with p53 in in vitro pull-down assays and in coimmunoprecipitation experiments of endogenous proteins (46). In this scenario, NS would function in the nucleoplasm, as p53 is not normally found in the nucleolus. Consistently, NS was found in the nucleoplasm at low levels and shown to rapidly shuttle between the nucleolus and the nucleoplasm (47).

To gain insight into the physiological function of NS, we generated a mouse line inactivated for NS. We show here that germ line inactivation of NS results in peri-implantation lethality between E3.5 and E5.5. Analyses of NS mutant blastocysts indicated that NS is not required to maintain pluripotency, nucleolar integrity, or survival of the pluripotent embryonic stem cells in vivo. However, proliferation of the mutant cells was decreased even in the absence of functional p53. Our results show that NS has a unique ability to control the proliferation rate of stem/progenitor cells in vivo independently of p53. Moreover, we show that this function of NS is evolutionarily conserved in early amphibian development.

MATERIALS AND METHODS

Generation of NS-deficient mice.

An embryonic stem (ES) cell clone with a single gene trap event in the NS locus was used to generate the NS-deficient mouse line (50). The vector used to create this particular ES cell clone was Rosaβ-geo, containing a splice acceptor site, the β-geo selection/reporter cassette, and a transcriptional stop element. To precisely locate the insertion site of the trapping vector, a PCR fragment was amplified, using primers at the 5′ end of the retroviral vector (5′-GGCCAGGATGAAGAGGCCTA-3′) and in the first exon of the NS locus (5′-GGCCAGGATGAAGAGGCCTA-3′), cloned, and sequenced. The insertion occurred in the first intron, 285 bp downstream of the 3′ end of exon 1 (Fig. 1A).

FIG. 1.

A gene trap event into the NS locus leads to a null NS mutation and early embryonic lethality in mice. (A) The structure of the retroviral construct along with the structures of the wild-type and mutated alleles and the positions of the oligonucleotides used for the PCR-based strategy are depicted. Exons are indicated by black boxes, and the long terminal repeats (LTR) of the viral vector are represented by light gray boxes. β-geo, lacZ/neomycin reporter/selection cassette. (B) PCR analysis of embryos with the indicated genotypes. The three primers, a, b, and c, are indicated. (C) Morphology of NS wild-type (+/+ or +/−) and homozygous mutant (−/−) E3.5 to E4.0 embryos. Magnification, ×200 (phase-contrast microscopy). (D) NS is expressed in the nucleoli of blastocysts. Blastocysts were recovered from NS^+/− intercrosses, fixed, and stained with anti-NS antibody (red). Nuclei were visualized with DAPI staining (blue). About 25% of the embryos failed to stain for NS. Magnification, ×200.

PCR genotyping.

A PCR-based strategy was developed to distinguish between the wild-type and NS mutant alleles. The primers are as follows: a, 5′-AACCGAGCCAGATTTCTGTG-3′; b, 5′-AACCGTCCCTGCAAAGTTATG-3′; c, 5′-GTCCTCCGATTGACTGAGTC-3′ (Fig. 1A). A 200-bp fragment indicates the presence of the mutated allele, whereas a 350-bp fragment is amplified from the wild-type allele (Fig. 1B). Genotyping of the p53-null and LSL alleles were determined as previously described (19, 17).

MEF preparation, culture, and BrdU incorporation.

Mouse embryonic fibroblasts (MEF) were prepared from E13.5 embryos and cultured as described previously (25). The procedure used for growth curves was previously described (25). The lifespan of MEFs was assayed by plating 1 × 10⁶ cells on 90-mm dishes in quadruplicate and passaging them on a 3T3 protocol. For bromodeoxyuridine (BrdU) incorporation assays, the MEFs were grown on 12-mm coverslips and pulse labeled with 10 μM BrdU for 1 h before fixation with 4% paraformaldehyde (PFA). The BrdU-positive cells were identified by indirect immunofluorescence.

Western blot analysis.

Lysates were prepared as previously described (25). Fifty micrograms of each sample was fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane. The primary antibodies used were anti-NS_mouse (rabbit polyclonal, 1/2,000; Chemicon) and γ-tubulin (mouse clone GTU-88, 1/7,500; Sigma) for normalization. The secondary antibodies were anti-rabbit-horseradish peroxidase (donkey polyclonal, 1/10,000; Amersham) and anti-mouse-horseradish peroxidase (sheep, 1/10,000; Amersham). For the detection, ECL Western blotting detection reagents (Amersham) were used.

Blastocyst collection, culture, and BrdU incorporation.

E3.5 blastocysts were flushed from the uterus and either used for immunofluorescence analysis or cultured in ES cell medium-MEF medium with added leukemia inhibitory factor. For the BrdU incorporation assays, blastocysts were cultured for 16 h in ES cell medium containing 50 μM BrdU (Sigma) before fixation in PFA.

Immunofluorescence.

Blastocysts or MEFs were fixed in 4% PFA for 15 min, permeabilized with 0.1% Triton, blocked with 1% bovine serum albumin (BSA), and incubated for 30 to 90 min with antiserum containing 1/100 anti-NS_mouse (Chemicon) plus one of the following: anti-BrdU (BU33; Sigma), anti-phosphorylated histone H3 (pHH3; Calbiochem), anti-Oct 3 (BD Transduction Labs), antifibrillarin (Abcam), or anti-cleaved caspase 3 (Cell Signaling). After another 30-min incubation with a secondary antibody (anti-mouse/rabbit or goat Alexa 488 or 633; Invitrogen/Molecular Probes), the embryos were mounted on slides using 4′,6′-diamidino-2-phenylindole (DAPI)-containing Vectashield mounting medium (Vector Laboratories).

Xenopus embryos and injections, RT-PCR, in situ hybridization, and histology.

Xenopus laevis eggs were obtained from hormone (chorionic gonadotropin; Sigma)-induced adult female frogs and fertilized using standard methods. Antisense morpholino oligonucleotides (MOs) for nucleostemin (Genetools) consist of the following sequences (the sequence complementary to the predicted start codon is underlined): Xns MO, 5′-TCCTTAGCTTCGGACGTTTCATGGC-3′; Xns-mis MO, 5′-TCATTAGCTTGCGACATTTCGTGGC-3′. MOs were injected at the two-cell stage unilaterally at 40 ng/blastomere. Injected embryos were fixed in MEMFA (MOPS [morpholinepropanesulfonic acid], EGTA, MgSO₄, formaldehyde), stained for β-galactosidase activity with 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (Bioline), and stored in ethanol at −20°C. Whole-mount phospho-histone H3 staining was performed essentially as described elsewhere (37). Whole-mount in situ hybridizations were performed using digoxigenin-labeled antisense RNA probes. The XNS probe was generated using T7 and Sal1-linearized pCMV-Sport6-XNS as templates (expressed sequence tag BC045248). Plasmids used for generating the other in situ probes were as described previously, FoxD3 (38) and N-tubulin (6). After completion of the whole-mount procedure, embryos were gelatin-embedded and vibratome-sectioned at a 30-μm thickness. RNA isolation from embryos was performed using the QIAGEN RNeasy minikit. RNA preparations were treated with DNase I (QIAGEN) and checked with 32 cycles of histone H4 for DNA contamination. Reverse transcription (RT)-PCR was carried out using the Gene Amp RNA PCR kit (Perkin-Elmer). The following primers were used: histone H4 F,5′-CGGGATAACATTCAGGGTATCACT-3′, and R, 5′-ATCCATGGCGGTAACTGTCTTCCT-3′ (58°C, 26 cycles); XNS F, 5′-CGAGAGACAGCGGCCATGAAA-3′, and R, 5′-CTTTTCAGAATTGTTATTTTTCTTCT-3′ (58°C, 30 cycles). PCR products were separated on 2% agarose gels and visualized by staining with ethidium bromide.

RESULTS

A gene trap event in the mouse NS locus results in peri-implantation lethality.

An ES cell clone (W223E05) with a specific gene trap event in the NS locus was microinjected into host blastocysts to produce an NS gene trap mouse line (50). The insertion gene trap vector integrated into intron 1 (Fig. 1A). This vector (Rosaβgeo) contains a splice acceptor (SA) site, the β-geo selection/reporter cassette, and a transcriptional stop element. The mutant allele is expected to drive the expression of a fusion protein that encodes the first four amino acids of the NS open reading frame and the β-galactosidase reporter. Notably, expression of the fusion transcript is under the control of resident NS regulatory element. Heterozygous mice for this mutation are viable and fertile. The ratio of wild-type to heterozygous mice indicated that loss of one copy of the NS gene was unlikely to cause developmental defects. However, matings between heterozygous animals produced no viable homozygous mutant offspring, indicating a recessive lethal phenotype (Table 1).

TABLE 1.

Genotyping analysis of progeny from NS heterozygous intercrosses

Stage	No. of litters	No. of embryos
		Total	+/+	+/−	−/−	Abnormal
E3-3.5	13	151	NDa	ND	43b (28)	43 (28)
E3-3.5	7	60	14	30	16 (26)	16 (26)
E5.5-7.5	11	97	ND	ND	ND (0)	5 (5)
E9.5-10.5	7	45	19	26	0 (0)	0 (0)
E12.5-13.5	6	57	17	40	0 (0)	0 (0)
E15.5-17.5	2	14	8	6	0 (0)	0 (0)
F2	49	342	127	215	0 (0)	0 (0)

^aND, not determined.

^bAs determined by lack of immunoreactivity to the NS antibody.

No homozygous mutants could be identified at E9.5 or later stages. We next performed serial sectioning and staining with hematoxylin and eosin of 11 litters at E5.5 or 7.5. Most embryos were morphologically normal at both stages, but five lacked any discernible embryonic tissue within the deciduae and so were considered presumptive NS-null embryos (Table 1). These presumptive null embryos showed complete lack of the development of the embryo proper, suggesting that the longest-surviving NS^−/− embryos die shortly after implantation.

At E3.5, blastocyst embryos are composed of an outer layer of cells that form an epithelium, the trophoectoderm, which gives rise to the trophoblast components of the placenta. The cells in the interior, the inner cell mass (ICM), develop into pluripotent progenitors of nontrophoblast extraembryonic tissues and all fetal cell types, including germ cells. At E3.5 to E4.5, all homozygous mutants showed various degrees of morphological deterioration and were distinguishable from their littermates. Whereas the NS^+/+ and NS^+/− embryos had gone on to form compact inner cell masses and spreading trophoblasts, NS mutants failed to form fully expanded blastocysts and showed reduced cell masses (Fig. 1C). At this stage of development, NS mutants could be identified by PCR (Fig. 1B, representative genotype assay) and a correlation between embryos showing a deteriorating phenotype and the NS^−/− genotype was observed. The mutant embryos could also be identified by immunofluorescence, which could readily detect NS protein expression in cells from both the ICM and trophoectoderm in wild-type blastocysts. In contrast, NS mutants are not immunoreactive to the anti-NS polyclonal antibody (Fig. 1D). Thus, this NS gene trap mutation leads to the absence of detectable levels of NS protein expression and results in peri-implantation lethality. Moreover, the NS^−/− embryos never hatched out of their zona pellucida during the study, but all of the wild-type and heterozygous embryos hatched at about E4.5 (data not shown).

The proliferation rate of NS^−/− ES cells is reduced in vivo.

NS mutant embryos appeared smaller than wild-type and heterozygous littermates. This could reflect a reduced number of cells in mutant embryos. To address this in a quantitative manner, E3 to 3.5 embryos from heterozygous intercrosses were collected, fixed immediately, stained for NS, and counterstained with DAPI, so that the number of nuclei and their morphology could be ascertained (Fig. 2A). At E3, NS^−/− embryos contained about half as many nuclei as the NS^+/− or NS^+/+ embryos. Moreover, whereas the number of cells in NS-immunoreactive embryos almost doubled between E3 and E3.5, the number of cells in the homozygous mutants showed no significant increase (Fig. 2A and B).

FIG. 2.

Failure of the NS^−/− blastocyst cells to proliferate normally. (A and B) number of nuclei detected by DAPI staining of blastocysts from NS^+/− intercrosses. (A) Mean cell number of wild-type and heterozygous embryos combined and of homozygous NS embryos at E3.0 and E3.5. Error bars denote standard errors. There is a significant difference between the mean numbers by Student's t test at both stages of development (E3.0, P = 6.10⁻¹³; E3.5, P = 1.10⁻¹¹). (B) Representative images of E3.5 NS^+/+ or NS^+/− and NS^−/− mutant embryos stained with DAPI (blue) and anti-NS (red) antibodies. Magnification, ×200. (C and D) Number of cells positive for phospho-Ser-histone H3 (P-H3, the M-phase specific marker) among the total number of blastocyst cells (determined by DAPI staining of nuclear DNA). (C) The percentage of P-H3-positive cells in wild-type and heterozygous embryos combined and in homozygous NS embryos at E3.0. Error bars denote standard errors. There is a significant difference between the mean numbers by Student's t test (P = 0.03). (D) Representative images of E3.0 NS^+/+ or NS^+/− and NS^−/− mutant embryos stained with DAPI (blue), anti-P-H3 antibody (green), and anti-NS antibody (red). Magnification, ×200. (E and F) NS-deficient blastocysts fail to incorporate BrdU. Blastocysts were collected and cultured for 16 h in ES cell medium containing 50 μM BrdU. (E) The percentage of BrdU-positive cells among the total number of blastocyst cells (determined by DAPI staining of nuclei) in wild-type and heterozygous embryos combined and in homozygous NS embryos at E4.0. Error bars denote standard errors. There is a significant difference between the mean numbers by Student's t test (P = 2.10⁻¹⁰). (D) Representative images of E4.0 NS^+/+ or NS^+/− and NS^−/− mutant embryos stained with DAPI (blue), anti-BrdU antibody (green), and anti-NS antibody (red). Magnification, ×200. n, number of embryos analyzed.

Notably, all NS^−/− nuclei were morphologically normal (Fig. 1D and 2B). The formation of micro- or macronuclei is a morphological signature of mitotic errors and is often associated with mutations in mitotic checkpoint genes or genes involved in the G₂ DNA damage checkpoint (1, 23, 45, 48). The absence of abnormal mitotic figures in NS mutant embryos indicates that NS deficiency is not associated with a defect in mitosis. Alternatively, the reduced cell number in NS mutant embryos could be due to increased spontaneous cell death. However, the absence of nuclear condensation and fragmentation in the NS mutants, as judged by fluorescent DNA stains (Fig. 2B), is not consistent with this possibility. This was further confirmed by measuring apoptosis directly by cleaved caspase-3 immunostaining. As shown in Fig. S1 in the supplemental material, NS mutant E3.5 to 4.5 embryos contained no more dying cells (Casp-3*-positive cells) than their wild-type or heterozygous littermates. Thus, NS deficiency does not cause loss of cell viability.

We next examined the potential effect of the NS^−/− mutation on the ability of the embryos to proliferate and progress properly through the cell cycle. Incorporation of BrdU into the DNA during the S phase of the cell cycle was assessed by indirect immunofluorescence. Both NS^+/+ and NS^+/− blastocysts showed strong incorporation of this DNA label, whereas the NS mutants showed practically no BrdU incorporation (Fig. 2E and F). This observation suggests that the proliferation capability of the cells from NS^−/− mutant embryos is significantly impaired at E3.0 to 3.5. We next immunostained E3.0 blastocyst embryos with both antibodies specific for NS and for histone H3 phosphorylated on serine residue 10 (P-H3). Phosphorylation of histone H3 at Ser 10 coincides with the initiation of chromosome condensation and is considered a specific marker of mitosis (42). Our analysis shows that the percentage of P-H3-positive cells is significantly lower in blastocyst cells from the NS mutant embryos than that in their wild-type or heterozygous counterparts (Fig. 2C and D). Together, these observations suggest that the majority of cells from NS-null blastocysts are arrested at E3.0 to 3.5 because the mutant cells fail to enter the S phase of the cell cycle.

The requirement of NS for proper cell proliferation might be linked to a role of this protein in the maintenance of nucleolar morphology and/or function. To test whether NS is required to maintain nucleolar integrity, blastocysts embryos were immunostained with antibodies directed against fibrillarin, a specific nucleolar marker. The fibrillarin staining was comparable in NS-null embryos and control littermates (Fig. 3A), indicating that NS deficiency does not cause nucleolar disruption.

FIG. 3.

NS is not required for nucleolar integrity or to maintain pluripotency. (A) Representative images of E3.5 NS^+/+ or NS^+/− and NS^−/− mutant embryos stained with DAPI (blue), antifibrillarin (green), and anti-NS (red) antibodies. Fibrillarin nucleolar localization is not perturbed in NS-deficient blastocysts. (B) Representative images of E3.5 NS^+/+ or NS^+/− and NS^−/− mutant embryos stained with DAPI (blue), anti-Oct3/4 (green), and anti-NS (red) antibodies. NS-deficient blastocysts maintain expression of Oct4, the marker of pluripotent stem cells. Magnification, ×200.

NS mutant cells might exit from the cell cycle because of spontaneous differentiation, accompanied with loss of pluripotency in the ICM. To clarify the fate of NS mutant cells, we immunostained the blastocysts with antibodies directed against Oct4 (also known as Oct3), a POU family transcription factor specifically expressed in ES cells, preimplantation embryos, epiblast, and germ cells (29, 40). Loss of Oct4 in blastocysts leads to terminal differentiation of the inner cell mass into trophoblast lineages (26), and precise Oct4 dosage is required for cell fate decisions (28). As expected, strong Oct4 staining could be detected in the ICM of wild-type and heterozygous NS embryos. Similar staining was observed in NS mutant cells (Fig. 3B), indicating that NS is not required for pluripotency in ICM.

Loss of p53 does not rescue NS deficiency.

NS is reported to be a p53-interacting protein, and expression of NS in stem cells was proposed to impose a block on p53 activity (5, 46). We, therefore, sought to test whether the embryonic lethality associated with NS loss depends on the presence of functional p53. NS heterozygous mice were crossed either with p53-null mice (19) or with mice homozygous for a p53 knock-in allele (p53^LSL/LSL) harboring a transcriptional stop element, flanked by LoxP recombination sites, and located in the first intron (A. Ventura and T. Jacks, unpublished data). This element renders the p53 locus transcriptionally silent, but p53 expression can be recovered in cells in vitro or tissues in vivo by expressing the Cre recombinase. NS^+/− p53^−/− mice were crossed with NS^+/− p53^LSL/LSL mice so that all animals produced had one p53-null allele and one p53^LSL allele. We had shown that this combination of alleles (p53^LSL/−) fully rescues the embryonic lethality associated with mdm2 and/or mdm4 loss (17). In contrast, loss of p53 expression did not rescue NS deficiency, because no viable NS^−/− p53^LSL/− mice were found (Table 2). Moreover, NS^−/− p53^−/− blastocysts were also morphologically distinguishable from their NS wild-type or heterozygous counterparts, which indicates that loss of p53 does not even partly prevent or delay the phenotype associated with NS loss (Fig. 4 A and B). Thus, this observation suggests that the ability of NS to control the proliferative potential of embryonic stem cells in vivo is not dependent on the growth suppressive activity of p53 and is not consistent with a role for NS upstream of p53.

TABLE 2.

Loss of p53 does not rescue NS deficiency^a

Stage	No. of litters	No. of embryos
		Total	NS+/+	NS+/−	NS−/−	Abnormal
E3.5	1	9	NDb	ND	2 (22)	2
E9.5	1	8	5	3	0 (0)	0
E12.5	6	45	13	32	0 (0)	0
F2	17	81	23	58	0 (0)	0

^aNS^+/− p53^LSL/LSL mice crossed with NS^+/− p53^−/− mice.

^bND, not determined.

FIG. 4.

p53 loss does not rescue NS deficiency; number of nuclei detected by DAPI staining of blastocysts from NS^+/− p53^−/− intercross matings. (A) Representative images of E3.5, p53-null, NS^+/+ or NS^+/− and NS^−/− mutant embryos stained with DAPI (blue) and anti-NS antibody (red). (B) Mean cell number of wild-type and heterozygous embryos combined and of homozygous NS embryos on a p53-deficient background at E3.0. Error bars denote standard errors. There is a significant difference between the mean numbers by Student's t test (P = 0.06). n, number of embryos analyzed. Magnification, ×200.

NS^+/− MEFs exhibit a reduced proliferative capacity.

Even if NS is found predominantly in undifferentiated cells, NS expression is not restricted to stem cells. Ubiquitous reporter (lacZ) expression in E10.5, E12.5, and E14.5 heterozygous embryos indicates that the NS promoter is active in virtually all embryonic cells at these stages of development (data not shown). Consistently, NS protein expression could be detected at high levels in total extracts from E10.5 embryos and in the nucleoli of explanted early passage MEFs (Fig. 5A and B). Interestingly, NS expression diminishes concomitantly with a decrease in the proliferation potential of the cells upon passages (Fig. 5A). Strikingly, expression of NS was consistently and significantly reduced in NS^+/− cells compared to NS^+/+ cells (Fig. 5A). To further study the implication of NS in the control of cell proliferation, we examined the proliferative capacity of the NS^+/− cells at various passages. We did not observe significant differences in the proliferation rate between heterozygous and wild-type cultures at early passages. However, at later passages, NS^+/− cells exhibited a slight but reproducible decrease in their proliferation potential compared to control cells (Fig. 5C). Together, the data suggest that NS is haplosufficient in the control of cell proliferation. Wild-type MEFs undergo 10 to 12 population doublings in culture, after which they cease to divide and develop a so-called senescent phenotype. Notably, even if a slight difference in the proliferation rate between heterozygous and wild-type cultures was observed on a 3T3 schedule, both NS^+/− and NS^+/+ cells underwent senescence at similar passages (Fig. 5C).

FIG. 5.

Haploid insufficiency of NS in MEFs leads to reduced proliferative potential. (A) NS endogenous protein is detected by Western blotting analysis in total extracts from E10.5 embryos (top panel) or in MEFs at various passages (lower panel). A significant decrease in NS protein expression is detected in cells prepared from NS^+/− heterozygous embryos, and NS expression decreases at late passages. γ-Tubulin (γ-tub) serves as a loading control here. (B) The NS protein (green) is mainly localized in the MEF nucleoli, as assessed by indirect immunofluorescence. Nuclei are visualized using DAPI staining. (C) Growth curves for NS^+/− and NS^−/− MEF cultures at passage 4. Cells (1 × 10⁵) were plated into six-well plates. Cultures were harvested at daily intervals, and the total number of cells was determined and normalized to the number of cells at day 1 (20 h after plating). The numbers refer to the mean values of four independently derived MEF cultures of each genotype. (D) Proliferation of NS^+/− and NS^−/− MEF cultures on a revisited 3T3 schedule. The accumulated number of population doublings is shown on a linear scale on the y axis.

NS has an evolutionarily conserved function in the control of cell proliferation.

NS is conserved during the evolution (see Fig. S2 in the supplemental material, alignment of the predicted amino acid sequences of NS from Drosophila, Caenorhabditis elegans, zebra fish, Xenopus, chicken, mouse, and human). We next used Xenopus laevis as a model system to further examine NS function in early embryos. We first determined the expression profile of XNS, the ortholog of NS in Xenopus laevis. XNS transcripts can be detected by RT-PCR analysis both maternally and zygotically (Fig. 6A -a). Whole-mount in situ hybridization shows that the entire neural plate of early neurula (stage 13) contains a significant amount of NS RNA, particularly around the anterior neural plate (Fig. 6A-b and -c). Cross-sections reveal that XNS is expressed throughout the ectoderm at the neurula stage and is restricted to the sensorial layer. By late neurula to tadpole stage, strong staining is observed in the head region in the brain, cranial neural crest cells, eye and otic placodes (Fig. 6A-e, -f, and -g). Weaker staining is observed in the somites. Expression of NS in the Xenopus embryos thus correlates with sites of active cell proliferation (37) and is therefore consistent with the above-described expression and function of NS in the mouse embryo.

FIG. 6.

XNS is expressed during early embryogenesis in proliferating tissues and its depletion affects the pattern of cell division and the expression of neuronal and neural crest markers. (A, part a) Expression of XNS during embryogenesis analyzed by RT-PCR. Control RT-PCR of histone H4 was used as a loading control. Note that XNS is expressed both maternally and zygotically. (A, parts b to i) Whole-mount in situ hybridization of XNS expression at the indicated stages Nieuwkoop-Faber stages (St) (27). (A, parts b and c) Dorsal views, anterior right. (A, parts d and i) Transversal sections at the indicated levels of the embryo shown in parts c and h, respectively. (A, parts e, f, and h) Lateral views, anterior right. (A, part g) Horizontal section of the embryo in part f at the level indicated. Note that XNS expression is first detected at the neurula stage in ectodermal cells of the sensorial layer, particularly around the anterior neural plate. At the tailbud and tadpole stages, XNS staining is observed in migratory cranial neural crests, the eye, and the otic vesicle. (i) Within the neural tube, XNS expression is restricted to the ventricular zone. np, neural plate; nc, neural crest; sle, sensorial layer of the ectoderm; ev, eye vesicle; s, somites; ov, otic vesicle; ba branchial arches. (B, part a) The XNS MO inhibits translation of XNS in vitro, whereas the XNSmis MO does not. (B, part b) Injection of the XNS MO, but not XNSmis, decreases the number of mitotic cells, as revealed using an antibody recognizing phosphorylated histone H3 (α-pH 3). Dorsal view of XNS MO- and XNSmis MO-injected embryos are shown. LacZ was coinjected as a lineage tracer. Control and injected sides revealed by LacZ staining are indicated (bottom panel). (d) Quantification of the results (percentage of pH 3-positive cells compared to control embryos). There is a significant difference between the percentages in control and MO NS-injected embryos by chi-square test (P = 0.002). n, number of cases analyzed. (C, parts a and d) XNS MO, but not XNSmis MO, decreases the expression of the indicated neuronal and neural crest markers at the neurula stage. The injected side is revealed by LacZ staining.

We next determined the requirement of XNS for proper neural plate formation and neural crest development. To this end, we designed MOs to specifically deplete the NS protein. We first determined that our NS MOs effectively block translation in an in vitro transcription/translation assay (Fig. 6B-a). Next, the NS and NS mismatch MOs were injected, together with β-galactosidase mRNA, into single cells of two-cell-stage embryos. This strategy allows us to knock down NS in the neural plate- and neural crest-forming regions of the ectoderm on one side of the embryo, with the other side serving as an internal control. Injection of the NS MOs, but not of the NS mismatch MOs, significantly decreases the number of mitotic cells in the injected side, as detected using an antibody recognizing phosphorylated histone H3 (Fig. 6B-c and -d). Moreover, NS MO-injected embryos display a decreased expression at the neurula stage of some neuronal (N-tubulin) and neural crest markers (FoxD3) (Fig. 6C-b and -d). However, expression of these markers is not completely abolished, because they were detected at later stages of development (data not shown). Together, these observations indicate that depletion of NS in Xenopus embryos only slows down neurogenesis, due to decreased cell proliferation, rather than interfering with neural specification. These data, therefore, further support a role of NS in the control of stem/progenitor cell proliferation and suggest that this function is conserved during the evolution.

DISCUSSION

NS was isolated because it is predominantly expressed in rat neuronal stem cells (46). It is also expressed in other stem cell-enriched populations, including embryonic stem cells. Consistent with this finding, we show herein that NS protein is detectable in mouse pluripotent stem cells in vivo. NS was found in E3 to E3.5 embryos both in the ICM and trophoectoderm. Interestingly, NS became undetectable when differentiation in CNS rat stem cells was induced (46). This decrease in expression is likely to reflect a decrease in the cell proliferation potential rather than being a consequence of cellular differentiation. NS is indeed not exclusively expressed in stem/progenitor cells. NS mRNA expression was, for instance, found in proliferating diploid fibroblasts or T lymphocytes (16). We also show here that NS protein is readily detectable in early passage MEFs but not at later passages when cells become senescent. Moreover, we generated an inactivated NS allele in mice and showed that NS^−/− blastocyst embryos retained expression of Oct-4, a marker of pluripotent stem cells in vivo. NS expression therefore appears closely associated with cellular proliferation, regardless of the cell of origin or its differentiation status, and it is not required for the maintenance of pluripotency. In the Xenopus embryo, NS expression was also predominant at various sites of intense cell proliferation. Interestingly, strong expression was detected in migratory cranial neural crest cells, a population of proliferative, migratory, tissue-invasive stem cells in a pattern roughly reminiscent to the one of Myc (4).

Previous observations supported an active role for NS in the control of cellular proliferation. Upon differentiation of neuronal stem cells, NS expression decreases before cell cycle withdrawal, suggesting that NS down-regulation leads to cell cycle exit rather than occurring as a consequence of the cell proliferation block. Moreover, when bone marrow stem cells are stimulated with fibroblast growth factor 2 to proliferate, NS expression increases in a dose-dependent manner, and knockdown of NS abolishes the proliferative effect of fibroblast growth factor 2 (20). Finally, NS is also strongly expressed in a number of cancer cell lines, and NS knockdown in an osteosarcoma cell line (U2OS) increased the number of noncycling cells (46). The data reported in this paper are consistent with the above observations. We showed that the NS^−/− mouse embryos die very early in development, as the blastocysts fail to enter S phase. Additionally, several different attempts to generate NS^−/− ES cells proved to be unsuccessful (data not shown). NS^+/− heterozygote MEFs showed about a 50% decrease in NS expression, and their proliferation rate was slightly less than that of wild-type cells. Finally, NS depletion in the Xenopus embryos resulted in defective proliferation of neural progenitor cells. Together, the data suggest an important, evolutionarily conserved role for NS in the control of cell cycle progression at early developmental stages.

The molecular basis for the requirement of NS for cell cycle progression remains unclear. It was proposed that the ability of NS to modulate cell proliferation is dependent on a functional interaction with the p53 pathway (46). This interaction may occur at several levels. NS might be required for nucleolar integrity and therefore for the regulation of p53 abundance (36). NS may affect the nucleoplasmic-nucleolar trafficking of key modulators of the p53 pathway. Alternatively, NS might affect p53 localization and/or function via direct interaction (46). However, our data are not consistent with these possibilities. First, nucleoli appeared to be intact in NS-deficient embryos, suggesting that NS loss does not cause nucleolar disruption. More importantly, loss of p53 does not rescue the embryonic lethality associated with NS inactivation. Moreover, like NS^−/− embryos, NS^−/− p53^−/− E3.0 embryos can be distinguished morphologically from the NS^+/+ p53^−/− or NS^+/− p53^−/− embryos, indicating that loss of p53 does not even partly rescue NS deficiency. The data therefore argue for a p53-independent function of NS in the control of cell proliferation.

NS protein is predominantly found in the nucleoli and thus may be involved in rRNA processing and ribosome assembly. Early studies highlighted the existence of cross talk between ribosome biogenesis and cell cycle progression (2, 41). More recently, perturbation of the activity of BOP1, a conserved nucleolar protein involved in rRNA processing and ribosome biogenesis, caused a cell proliferation block in the G₁ phase of the cell cycle (43). Similarly, lack of BrdU incorporation of the NS-deficient embryos suggests that NS is required for the G₁/S phase transition. Moreover, NS belongs to a subclass of GTPases characterized by a circular permutation of their GTPase signature motifs from the C-terminal side of the protein to the N terminus (12). Several members of this family, including NGP-1, Nug1p, Nug2p/Nog2p, and Grn1p/GNL3L (34, 3, 39, 13), have been shown to localize to the nucleolus and participate in the processing of the nucleolar pre-rRNA, ribosomal assembly, and nucleolar/nuclear export. For instance, deletion of Grn1 resulted in a severe growth defect, a marked reduction in mature rRNA species with a concomitant accumulation of the 35S pre-rRNA transcript, and failure to export the ribosomal protein Rpl25a from the nucleolus (13). Thus, NS-deficient cells might therefore undergo cell cycle arrest because of a defect in ribosome biogenesis/rRNA processing. Mammalian cells quickly adjust the rate of ribosome synthesis based on the availability of nutrients and growth-promoting mitogens, and cells that exit the division cycle into a quiescent state greatly limit ribosome production and overall protein synthesis (15, 32). How ribosome biogenesis is ultimately coordinated with cell growth and proliferation remains unclear. Further experiments, using conditional knockout or knockdown approaches in mice and/or primary cells should be carried out to formally test the contribution of NS in these processes.

The observation that loss of NS function and even decreased NS protein levels impairs cell cycle progression has potential therapeutic implications. The presence in the protein of two putative GTP binding sites makes it a very attractive target for molecules that would mimic or compete for GTP binding. On the one hand, these molecules could be tested for their ability to affect cancer cell proliferation. On the other hand, as expansion of stem cells in culture is a key tool for tissue regeneration, drugs that potentiate NS activities may facilitate stem cell maintenance and propagation.