SEARCH

SEARCH BY CITATION

Keywords:

  • Cell culture;
  • Stem/progenitor cell;
  • Embryo;
  • Neural differentiation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References
  11. Supporting Information

Embryonic stem cells differentiate into neuroectodermal cells under specific culture conditions. In primates, these cells are organized into rosettes expressing Pax6 and Sox1 and are responsive to inductive signals such as Sonic hedgehog (Shh) and retinoic acid. However, direct derivation of organized neuroectoderm in vitro from preimplantation mammalian embryos has never been reported. Here, we show that bovine inner cell masses from nuclear transfer and fertilized embryos, grown on feeders in serum-free medium, form polarized rosette structures expressing nestin, Pax6, Pax7, Sox1, and Otx2 and exhibiting interkinetic nuclear migration activity and cell junction distribution as in the developing neural tube. After in vitro expansion, neural rosettes give rise to p75-positive neural crest precursor cell lines capable of long-term proliferation and differentiation in autonomic and sensory peripheral neurons, glial cells, melanocytes, smooth muscle cells, and chondrocytes, recapitulating in vitro the unique plasticity of the neural crest lineage. Challenging the rosette dorsal fate by early exposure to Shh induces the expression of ventral markers Isl1, Nkx2.2, and Nkx6.1 and differentiation of mature astrocytes and neurons of central nervous system ventral identity, demonstrating appropriate response to inductive signals. All together, these findings indicate that neural rosettes directly derived from cloned and fertilized bovine embryos represent an in vitro model of early neural specification and differentiation events. Moreover, this study provides a source of highly proliferative neural crest precursor cell lines of wide differentiation potential for cell therapy and tissue engineering applications.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References
  11. Supporting Information

During the process of neurulation in vertebrate embryos, the ectoderm folds at its most dorsal point, giving rise to the outer epidermis and the inner neural tube while the neural crest originates in between. In mammals, in vitro models of neural differentiation have been reported for embryonic stem cells (ESCs), both mouse and primate, whereby differentiation of ESCs in neuroectodermal cells generally occurs in serum-free medium or in coculture with stromal feeders that produce inductive signals [1, [2]3]. Under such conditions, primate ESCs readily organize into rosette structures resembling the developing neural tube [4, 5], although a direct comparison between neural rosettes and the embryonic neural tube has never been described. A number of neural cell types have been derived from ESCs, including midbrain dopaminergic neurons and spinal motoneurons [1, 5, 6]. More recently, a self-renewing population of neural stem (NS) cells has been obtained from ESCs after serum-free culture and expansion in the presence of basic fibroblast growth factor (bFGF) and epidermal growth factor (EGF) [7]. By contrast, minor attention has been given to the neural crest lineage, and only a few reports have demonstrated the derivation of cell types of neural crest identity from ESC cultures of both mouse and primate [3, 8].

The aim of this study was to devise a method of neural induction directly from a mammalian embryo, without the intermediate stage of ESC derivation, to provide a novel in vitro model of early neural differentiation events in mammals. This study is carried out in a bovine model, a species in which the conditions for true ESC derivation have not been established yet, although several reports have indicated that bovine inner cell mass (ICM) cells can be cultured in vitro and induced to differentiate in a variety of cell types [9, 10].

Our results show that organized neuroectoderm, in the form of neural rosettes, can be directly derived from both fertilized and cloned bovine embryos. We demonstrate that neural rosettes display several morphological and functional features typical of the developing neural tube. Moreover, after in vitro expansion of rosette cells with growth factors, we characterized a process reminiscent of the emigration of neural crest cells from the neural tube, giving rise to highly proliferative neural crest precursor cell lines capable to differentiate in all neural crest derivatives.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References
  11. Supporting Information

Unless otherwise indicated, all chemicals were purchased from Sigma-Aldrich (St. Louis, http://www.sigmaaldrich.com).

Production of Cloned and Fertilized Bovine Embryos

The procedure for generating bovine embryos by in vitro maturation and nuclear transfer (NT) or by in vitro fertilization (IVF) was described previously [11]. Briefly, bovine oocytes, recovered from abattoir ovaries, were matured in TCM199 supplemented with 10% (vol/vol) fetal calf serum, ITS media supplement (1 μl/ml), 1 mM sodium pyruvate, 0.5 mM l-cystein, 10 mM glycine, 100 μM β-mercaptoethanol, 50 ng/ml long-EGF, 100 ng/ml long-insulin-like growth factor-1, and 10 ng/ml bFGF at 38.5°C in 5% CO2 in humidified air for 18–22 hours. For NT embryo construction, the zona pellucida was removed by pronase digestion [12], and oocytes were enucleated under UV light, fused with serum-starved adult skin fibroblasts and activated with 5 μM ionomycin for 4 minutes followed by 4 hours of culture in 2 mM 6-dimethyl-amino-purine. Finally, the embryos were transferred in synthetic oviduct fluid supplemented with essential and nonessential amino acids (SOFaa) medium [13] with 16 mg/ml bovine serum albumin in 5% CO2 and 5% O2 up to day 7–8 (fertilization = day 0). For production of IVF embryos, matured oocytes were fertilized in SOFaa medium without glucose supplemented with 1 μg/ml heparin, 20 μM d-penicillamine, 100 μM hypotaurine, and 1 μM epinephrine. Motile spermatozoa, obtained by centrifugation of frozen thawed semen on Percoll discontinuous density gradient, were added at a final concentration of 0.5 million sperm per milliliter. After incubation for 18–20 hours, presumptive zygotes were denuded of cumulus cells by vortexing and were transferred in SOFaa medium as described above.

Derivation of Neural Rosettes and Neural Crest Precursor Cell Lines

On day 7–8, ICMs of blastocyst stage embryos, both NT and IVF, were isolated with insulin needles or by immunosurgery and plated on mitomycin (Sigma-Aldrich)-inactivated STO fibroblasts in Dulbecco's modified Eagle's medium (DMEM)/F-12 medium supplemented with 15% knockout serum replacement (Invitrogen, Carlsbad, CA, http://www.invitrogen.com), 2 mM glutamine, 100 μM β-mercaptoethanol, with or without 2 ng/ml bFGF. One week later, the outgrowths were disaggregated with 0.05% trypsin-EDTA and replated. After 10 days, rosettes containing colonies were separated from STO feeders, cut into small fragments with insulin needles, and plated on matrigel (1:100 dilution)-coated dishes in DMEM/F-12 medium supplemented with 0.6% glucose, 3 mM sodium bicarbonate, 2 mM glutamine, 5 mM HEPES, 25 μg/ml insulin, 60 μM putrescine, 20 nM progesterone, 100 μg/ml transferrin, 30 nM sodium selenite, 2 μg/ml heparin, 10 ng/ml bFGF, and 20 ng/ml EGF. After 2–3 days, the cultures were trypsinized in single cells, plated at 20,000 per cm2, and further passaged at 3–4-day intervals on matrigel-coated dishes.

In Vitro Differentiation

Neural crest precursor cell lines were differentiated by growth factor withdrawal and supplementation of ascorbic acid (200 μM) starting from passage three onward, approximately every five passages. The cultures were fixed in 4% paraformaldehyde (PFA) 10–35 days after induction.

In Vivo Differentiation

From 40 to 70 million neural precursor cells at passages 3–4 or at passages 15–20 derived from four different embryos were injected subcutaneously in nonobese diabetic-severe combined immunodeficient (NOD-SCID) mice. Animals were sacrificed 4 and 9 weeks later for early passage and late passage transplants, respectively. Tumors were fixed in 4% PFA for immunohistochemistry.

Bromodeoxyuridine Staining

Rosettes were incubated in medium containing bromodeoxyuridine (BrdU) (3 μM, 30 minutes) before fixation in 4% PFA followed by postfixation in 70% ethanol. To perform immunohistochemistry against BrdU, DNA was chemically denatured by incubation in 2 N HCl for 30 minutes. Pregnant female CD1 mice at embryonic day 10.5 were injected with BrdU (100 μg/g of body weight) and sacrificed 90 minutes later. Immunohistochemical stainings on embryos were performed as previously described [14].

Immunohistochemistry and Microscopy

Cells were fixed with 4% PFA with or without 0.2% gluteraldehyde. Unspecific binding blocking (10% donkey serum, 0.1% Triton in phosphate-buffered saline [PBS]) was followed by primary antibody incubation (4°C, overnight) after which secondary antibody incubation (1 hour at room temperature) was performed. Finally, slides were mounted in mounting medium (DakoCytomation, Glostrup, Denmark, http://www.dakocytomation.com). Primary and secondary antibodies are listed in the supplemental online data. For histology, slides were hydrated and stained for 5 minutes with Alcian Blue (1% in 3% CH3COOH water solution), toluidin blue (0.02% in H2O), or hematoxylin (0.75%) and eosin (1%).

Fluorescence-Activated Cell-Sorting Analysis

One million cells were trypsinized, washed in PBS twice, and incubated with either no primary antibody or phycoerythrin (PE)-conjugated anti-human p75 (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com) or an isotype control (15 minutes at 4°C). Cells were then washed and analyzed on a Beckman Coulter (Fullerton, CA, http://www.beckmancoulter.com) Cytomics FC500 flow cytometer. PE fluorescence resulting from direct excitation at 488 nm by the argon laser was detected using a 575-nm bandpass filter and compared with the isotype control. A minimum of 20,000 cells was analyzed for each sample.

Reverse Transcription-Polymerase Chain Reaction Analysis

Total RNA was extracted from bovine blastocysts, multiple rosette colonies, or undifferentiated neural crest precursors (one million) with TRIzol reagent (Invitrogen). cDNA was synthesized using Thermoscript reverse transcription-polymerase chain reaction (RT-PCR) system (Invitrogen) according to the supplier's protocol and was used as the template for PCRs. Reactions were performed in 25 μl containing cDNA, primers (as listed in supplemental online data), one x PCR master mix, and AmpliTAQ Gold (Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com).

All experiments that involved the use of animals were carried out under veterinary supervision in accordance with Decreto Legislativo 116/92—which regulates the use of animal experimentation in Italy—and were approved by the Local Ethical Committee of Laboratorio di Tecnologie della Riproduzione (Cremona, Italy).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References
  11. Supporting Information

Derivation of Neural Rosettes

Isolated ICMs from fertilized and cloned bovine embryos were plated on inactivated STO feeders in serum-free medium with or without bFGF supplementation. The plating efficiency was very high, approaching 90% (35/40). One week after plating, the derived outgrowths, still positive for the totipotency markers Oct4 and Sox2, were disaggregated, and after an additional 10 days of culture, extensive formation of colonies containing clusters of rosettes was observed (Fig. 1A–1D) in approximately 60% of the outgrowths, with no difference between NT and IVF embryo outgrowths and both in the presence and absence of exogenous bFGF.

thumbnail image

Figure Figure 1.. Derivation and characterization of neural rosettes. (A): Day 7 bovine blastocyst. (B): Primary outgrowth. Rosettes colony on feeders (C), detached from feeders (D), and on matrigel-coated dish (E). Rosettes stained positive for nestin, Otx2, Pax6, and Pax7(F–I). Reverse transcription-polymerase chain reaction on blastocysts and rosettes. Bovine blastocysts are positive for the totipotent markers Oct4, Rex1, and Sox2. Neural rosettes maintain Sox2 but downregulate Oct4 and Rex1 and upregulate Sox1 together with a panel of forebrain (Otx1, Pax6, and Emx2), midbrain (Otx1 and Dmbx1), and hindbrain (Hoxb1) genes, whereas spinal cord genes (Hoxb2, Hoxb4, and Hoxb9) are not expressed (J). Labeling of S-phase nuclei (BrdU staining) and mitotic nuclei (PH3 staining) in rosettes (K, L, N, O) and embryonic day 10.5 murine neural tube (M, P). Colocalization of BrdU and PH3 (Q–S). Apical localization of cell junctions stained with ZO-1 and β-catenin (adherens junctions) in rosettes (T, V) and neural tube (U, W). Peripheral localization of laminin in rosettes (X) and neural tube (Y). Scale bars = 50 μm (A–C, E, F, L, O, R, X), 100 μm (G–I, K, M, N, P, Q, S, W, Y), and 3 mm (D). (A–C, F–S) are derived from in vitro fertilized embryos; (D, E) are derived from nuclear transfer embryos. Abbreviations: Blast, blastocyst; BrdU, bromodeoxyuridine; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; PH3, phospho-histone H3; Roset, rosette.

Download figure to PowerPoint

Immunocytochemical analyses were performed on isolated rosette colonies after mild collagenase treatment and plating in the absence of feeders. Rosettes stained positive for markers of early neural tube such as nestin, Otx2, and Pax6 and for the dorsal marker Pax7 (Fig. 1F–1I). These findings were more extensively confirmed by RT-PCR, in which rosettes were found positive for Sox2, Sox1, and nestin and negative for Oct4, whereas Rex1 had lower expression as compared with intact blastocysts, indicating the loss of the ICM undifferentiated state. To identify their regional identity, rosettes were tested and found positive for the expression of markers of forebrain (Otx1, Foxg1, and Emx2), midbrain (Otx1 and Dmbx1), and hindbrain (Hoxb1), whereas spinal cord genes (Hoxb4 and Hoxb9) were not expressed (Fig. 1J). In addition, the expression of Pax7, but not of Nkx2.2 and Isl1, was detected, indicating that neural rosettes directly derived from bovine embryos, under the conditions described in this study, are dorsalized neural structures.

Neural rosettes derived in vitro are morphologically reminiscent of the developing neural tube in vivo; however, no direct comparison has been reported yet. A well-known feature of neural tube cells is the movement of their nuclei according to the cell cycle stage referred to as interkinetic nuclear migration [15]. Remarkably, we found that nuclei within rosettes undergo similar migration. Nuclei undergoing DNA synthesis (BrdU staining, S-phase labeling) were located at the outer edges (Fig. 1K–1M), whereas mitotic nuclei were always confined to the luminal side, as shown by phosphohistone H3 (PH3) antibody staining (G2- to M-phase staining) (Fig. 1N–1P) [16]. PH3 and BrdU stainings never overlapped, indicating a highly controlled spatial confinement, as normally occurs in the proliferating neural tube (Fig. 1Q–1S). In vivo neural tube cells are tightly connected by cell junctions. In particular, adherent junctions and some components of tight junctions, such as ZO-1 protein, are localized mainly at the apical side (Fig. 1U, 1W) [15, 17]. As expected, we found that rosettes were stained with β-catenin and ZO-1 antibodies on the apical cytoplasmatic side, indicating polarization of cell junctions (Fig. 1T, 1V). Moreover, a clear staining for laminin was observed at the periphery of the rosettes (Fig. 1X), as occurs in the basement membrane that surrounds the neural tube (Fig. 1Y). Finally, neural rosettes were analyzed for markers of neurogenic radial glia that in vitro has been described as a transient [18] or stable [7] stage of ESC neural differentiation and that in vivo is the precursor of both neurons and astrocytes during development of the nervous system [18, 19]. Rosette cells were found positive for Glast and vimentin, and Glast expression was detected also by RT-PCR together with brain lipid binding protein (supplemental online Fig. 6).

Characterization of Neural Crest Precursors

To further investigate the dorsal specification identified by RT-PCR, the rosette colonies were probed with the neural crest marker p75. Remarkably, we found extensive p75 staining mainly at the periphery of the rosettes, indicating a clear neural crest specification (Fig. 2A). After dissociation in single cells by trypsin treatment, the suspension of rosette cells was plated on matrigel-coated dishes in DMEM/F-12 medium supplemented with bFGF and EGF and additives commonly used for neural stem cells. The morphology of the derived cultures was predominantly as illustrated in Figure 2B, 2E (9/13 derivations, six from NT and three from IVF embryos). Due to the proliferation of the outer cells and the parallel reduction in size and cell number of the rosettes proper, the rosette structures were progressively lost with passage number (Fig. 2B). Immunocytological analyses confirmed that the proliferating cells were p75-positive, and Pax6 staining was observed in the residual inner rosette cells (Fig. 2C). With passaging, virtually all cells became positive to p75 as shown by immunofluorescence and fluorescence-activated cell-sorting analysis (Fig. 2E–2G). In parallel, RT-PCR showed a progressive loss of Sox1 and an increased expression of the typical neural crest marker Slug followed by Msx1, Sox10, and FoxD3, confirming a clear neural crest specification (Fig. 2D, 2H). These findings further support the hypothesis that rosettes display morphological and functional features common to the developing neural tube and also give rise to p75-positive cell lines through a mechanism that is reminiscent of the emigration of neural crest cells from the neural tube itself.

thumbnail image

Figure Figure 2.. Identification and isolation of neural crest precursors. (A, B): Cells at the periphery of rosettes (asterisks) were p75-positive. (B): With passaging, Pax6-positive rosette inner cells (arrows in [B, C], passage 3) were progressively lost ([E, F], passage 7). Similarly, Sox1 was downregulated (D), whereas neural crest-specific genes were upregulated (H). Fluorescence-activated cell-sorting analysis showed that more than 90% (n = 3) of the population express p75 above the background level represented by an isotype control. (I): Representative growth curve of one established neural crest precursor cell line. Scale bars = 50 μm (B, C, E, F) and 100 μm (A). (A, C–F, H) are derived from in vitro fertilized embryos; (B, G, I) are derived from nuclear transfer embryos. Abbreviations: GAPDH, glyceraldehyde-3-phosphate dehydrogenase; PE, phycoerythrin.

Download figure to PowerPoint

Four of the nine cell lines, three from NT and one from IVF embryos, characterized by the neural crest markers mentioned above were cultured up to 120 population doublings. The growth curve was linear (Fig. 2I), and the average doubling time was 22.8 ± 2.1 hours. Karyotype analysis showed that three out of four lines were karyotipically normal (n = 60). The remaining four cell lines (out of the 13 derivations) had a mixed morphology in which both migrating cells and rosette cells appeared to proliferate. In addition, they were found positive both for Sox1 and Slug (indicating a mixed phenotype), were capable of limited proliferation, and were not investigated further. By contrast, when the nine cell lines expressing neural crest markers were induced to differentiate by growth factor withdrawal and addition of ascorbic acid, a wide variety of cell types were characterized. Groups of β-III-tubulin-positive neurons together with large clusters of pigmented cells positive for Pmel17, a specific marker for mature melanocytes, were identified (Fig. 3A–3D). Further characterization revealed neurons positive for peripherin, most of which (75% ± 5%, n = 155) were found positive for Brn3A (Fig. 3E), a marker for the sensory lineage, whereas tyrosine hydroxylase identified autonomic neurons (15% ± 3%, n = 155), mainly found in clusters, and only a few were found isolated (Fig. 3F), therefore confirming their neural crest identity. Among the differentiated cells were patches of CNPase- (Fig. 3G), A2B5- (Fig. 3H), and S100- (not shown) positive cells, indicating differentiation toward an oligoglial phenotype, but we were unable to detect markers of mature Schwann cells. We also found large areas of smooth muscle actin-positive cells (Fig. 3I), again consistent with the derivation of a smooth muscle phenotype from neural crest precursor cell lines. When differentiation was prolonged for more than 2 weeks, nodular structures were formed consisting of bright cells surrounded by an acellular matrix. Such structures increased in size and number both with the time of differentiation and the passage number of the cell lines. Alcian Blue and collagen II staining demonstrated the cartilaginous nature of the nodules (Fig. 3J, 3K), and von Kossa staining indicated a variable degree of calcium accumulation (Fig. 3L). Thus, we could conclude that the full differentiation program of cranial neural crest, including neurons, melanocytes, glia, smooth muscle, and chondrocytes, was recapitulated in vitro starting from neural crest precursors derived from preimplantation bovine embryos.

thumbnail image

Figure Figure 3.. In vitro differentiation of neural crest precursors. (A, B): Differentiated cultures contained clusters of neurons (A) and patches of pigmented cells (B) positive to pmel17 (C). (D): General neural commitment and specific peripheric neuronal differentiation were identified by β-III-tubulin and peripherin stainings, respectively. Brn3a (E) and tyrosine hydroxylase (F) revealed either sensory or autonomic neuronal fate. Glial precursors were identified by CNPase and A2B5 staining (G, H), and smooth muscle cells were identified by smooth muscle actin staining (I). Nodule-like structures were positive for both Alcian Blue ([J], a 3.5-cm Petri dish at 30 days of differentiation) and collagen II ([K], histological section), indicating cartilage differentiation, whereas calcium deposition was revealed by von Kossa staining (L). Scale bars = 50 μm (E, F, I, K) and 100 μm (A–D, G, H, L). (C–H, K) are derived from in vitro fertilized embryos; (A, B, I, J, L) are derived from nuclear transfer embryos.

Download figure to PowerPoint

Neural Rosettes Respond to Inductive Signals

During neural tube morphogenesis, Shh suppresses the development of dorsal tissues, including neural crest formation, and promotes the differentiation of ventral central nervous system (CNS) tissues [20]. Therefore, we asked whether cell fate determination during rosette establishment in vitro is responsive to factors known to refine precursor domains inside the neural tube. For this purpose, we challenged the dorsal fate commitment of rosettes by exposure to Shh for 12 days after the first disaggregation of the outgrowths. We then withdrew all growth factors and analyzed the cells 14 days later. Shh-treated cells lacked Pax7, Slug, Sox10, and Msx1 expression but maintained Sox1 and activated a set of genes specifically confined to the ventral neural tube such as Nkx6.1, Isl1, and Nkx2.2 and in particular to the floor plate (FoxA2 and Shh) (Fig. 4A) [21]. When the Shh-treated cells were induced to differentiate, 72% ± 6% (n = 135) of the cells stained positive for β-III-tubulin (Fig. 4B) and neuronal cell adhesion molecule but not peripherin, indicating CNS neuronal commitment. Furthermore, glial fibrillary acidic protein-positive astrocytes were detected (7% ± 2%, n = 135) (Fig. 4C). The remaining cell population resulted vimentin+ and β-III-tubulin, suggesting an undifferentiated neural precursor state (not shown). Neurons were mostly GABAergic, (48% ± 8%, n = 125) and glutamatergic (32% ± 6%, n = 125) but not dopaminergic (Fig. 4D, 4E). Isl1+ and Nkx2.1+ neurons were found scattered on the plate, further supporting their ventralized fate (Fig. 4F, 4G).

thumbnail image

Figure Figure 4.. Sonic hedgehog (Shh) treatment. (A): Shh repressed central nervous system dorsal markers such as Pax7 as well as neural crest-specific genes like Slug, Sox10, and Msx1, whereas genes of the ventral neural tube (Nkx6.1, Isl1, and Nkx2.2) and the floor plate (FoxA2 and Shh) were induced. After differentiation, large numbers of β-III-tubulin (βIIItub)-positive neurons (72% ± 4%, n = 148) (B) and some scattered glial fibrillary acidic protein-positive astrocytes (8% ± 2%, n = 148) (C) were identified. No signs of neural crest differentiation were observed. β-III-Tubulin-positive neurons were identified as GABAergic (D) or glutamatergic (E) but not dopaminergic. Differentiated neurons were found to be positive for ventral-specific markers such as Nkx2.1(F) and Isl1(G). Scale bars = 50 μm (F, G) and 100 μm (B–E). (A–G) are derived from in vitro fertilized embryos. Abbreviation: GFAP, glial fibrillary acidic protein.

Download figure to PowerPoint

In Vivo Differentiation in Ectopic Sites

We finally tested the in vivo differentiation and tumorigenicity of neural crest precursor cell lines by in vivo transplant in an ectopic site. Cells at early (passages 3–4) and advanced passages (passages 15–20) were injected subcutaneously in NOD-SCID mice. Tumors were formed in four of six early transplants and four of four late transplants. Their weight ranged from 730 to 1,620 mg for early transplants and from 145 to 1,540 mg for late transplants. Early transplants gave rise to large soft tumors of uniform appearance composed exclusively of tubular/rosettes structures (Fig. 5A–5C) similar to the organization typically found in primitive neuroectodermal tumors [22]. Tumor cells stained positive for β-III-tubulin and vimentin (Fig. 5D, 5E). Some staining for p75 was also detected in between the rosettes/tubules (Fig. 5F).

thumbnail image

Figure Figure 5.. Subcutaneous injections of neural crest precursors in nonobese diabetic-severe combined immunodeficient mice. (A): Tumoral mass derived from subcutaneous injections of neural crest precursors at passages 3–4. (B–D): The growing tissue was organized in rosettes (B) expressing vimentin (C) and displaying features similar to their in vitro counterparts, such as ZO-1 staining (green) and phospho-histone H3-positive mitotic nuclei (D). (E): Cells between rosettes were identified mainly as β-III-tubulin-positive neurons. (F): Some p75-positive neural crest precursors were also identified. (G, H): Tumoral masses derived from subcutaneous injections of neural crest precursors at passages 15–20 of uniform, pearly appearance and hard consistency. (I): Hematoxylin-eosin staining revealed a homogenous cartilaginous organization of the tumoral tissue. (J): Alcian Blue staining of an area with arrayed and hypertrophic chondrocytes. (K): A well-developed perichondrium-like structure observed at the edges of the tissue (arrowheads). (L): Extracellular cartilage matrix characterized by metachromatic staining with toluidin blue. Scale bars = 12 mm (A), 100 μm (B–F), 10 mm (G, H), 300 μm (I, J), and 50 μm (K, L). (A–F) are derived from nuclear transfer embryos; (G–L) are derived from in vitro fertilized embryos.

Download figure to PowerPoint

These findings indicated that although rosette cells maintain a high proliferative ability in ectopic locations, their differentiation potential is restricted to neural fate. By contrast, tumors of hard consistency were derived from the more advanced passages (Fig. 5G, 5H). Cytological analysis with alcian blue (Fig. 5J) and toluidin (Fig. 5L) stainings indicated that the tumors were solid masses of cartilage surrounded by a well-differentiated perichondrium. Thus, cell fate restriction obtained by in vitro induction of rosette cells into neural crest progenitors was maintained in vivo also in ectopic graftings, suggesting the establishment of a stable cell fate identity. In no instance was teratocarcinoma formation observed.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References
  11. Supporting Information

In this paper, we have described the direct derivation of organized neuroectoderm in vitro, in the form of neural rosettes, from bovine blastocysts. The high efficiency of derivation, both from fertilized and NT embryos, indicates a strong neural commitment of bovine ICM cells under serum-free culture conditions as previously shown for mouse ESCs and human ESCs (hESCs). Interestingly, we found that serum (10%) or bone morphogenetic protein (BMP)-4 (10 ng/ml) supplementation completely blocked rosette formation in agreement with the inhibitory role that BMPs exert on neural induction in vivo [23, 24]. We also found that bFGF supplementation is not required for bovine neural rosette formation as previously shown for hESCs [5]. These findings suggest a remarkable analogy between primate ESCs and bovine ICM cells in the response to neural differentiation conditions. Moreover, our results are in overall agreement with the “neural default” model described in fish and amphibia, according to which ectodermal cells will undergo neural differentiation unless instructed by BMPs and related molecules to follow an epidermal cell fate. Thus, differentiation of bovine embryonic cells, using the method described in this study, represents a powerful in vitro model for investigating the molecular mechanisms of neural induction in mammals, in which results have not been conclusive so far.

Bovine neural rosettes express early neural marker genes such as nestin, Sox1, and Pax6 and are characterized by a peculiar spatial distribution of cell junctions and basal lamina proteins and by typical interkinetic nuclear movements never reported before for neural rosettes derived from primate ESCs. This latter morphological feature, typical of the developing neural tube in vivo, has been recently described in vitro by Conti et al. [7] for radial glia cell lines (NS cells) obtained from mouse ESCs and hESCs. In vivo, in the developing cerebral cortex, radial glia accounts for most cortical neural progenitors, producing most of the neurons and potentially those from many other CNS regions [25, 26], and for these reasons they have been considered as the precursors of the neural stem cells residing both in the subventricular zone and the dentate gyrus of the adult brain [27]. It is not surprising therefore that rosette cells in this study are positive for radial glia markers. Moreover, our results confirm that radial glia represent a common state, transient or stable, already described in the course of in vitro neural differentiation of ESCs [2, 7, 28] and reported here after direct derivation of neural precursors from the embryo.

During embryonic development, Shh is the principal ventralizing signal that patterns the developing neural tube. In this study, we showed that Shh could ventralize the otherwise dorsal fate of bovine neural rosettes, indicating that rosette cells respond to the same inductive signals inducing dorsoventral patterning during neurulation in vivo. Taken all together, these morphological and functional features indicate that rosettes derived from bovine embryos are indeed a close in vitro representation of early neurulation events in mammals.

We reported that, after enzymatic disaggregation, neural rosettes progressively give rise to a culture of p75-positive cells expressing neural crest-specific genes. Contrary to tissue-specific neural crest precursors, characterized by limited proliferation [29], the embryo-derived neural crest precursor cell lines described in this study are capable of more than 120 population doublings maintaining a normal karyotype. In vitro differentiation experiments across all of the derived cell lines have confirmed their neural crest identity: all neural and mesenchymal derivatives, including neurons, melanocytes, glia, smooth muscle, and chondrocytes, have been characterized in differentiated cultures. However, with each passage number, we observed a progressive reduction of neural derivatives to the advantage of ectomesenchymal derivatives. This observation was confirmed in our in vivo differentiation experiments in immunocompromised hosts, in which early passages of neural crest precursors gave rise to neuroectodermal tumors, and late passages gave rise to cartilaginous tumors only. This restriction of differentiation is likely a consequence of the deregulatory effect of prolonged exposure to bFGF during expansion [30] as already reported for neural precursors derived from hESCs [5]. Interestingly, NS cells [7] have also been derived in serum-free media supplemented with bFGF and EGF during a brief period of suspension culture followed by expansion of the only cell fraction adherent to the plastic. Under these conditions, the selected population maintained both a strong neuronal commitment and a high replication ability. These findings demonstrate that different culture and selection protocols influence the characteristics of neural precursors derived from ESCs, and preliminary experiments in our laboratory suggest that a similar strategy could be applied successfully to derive neural precursors with long-term neuronal differentiation ability directly from bovine embryos.

In has been shown that mouse and primate ESCs, cultured on the stromal cell line PA6, can be induced to differentiate in neurons and smooth muscle cells of neural crest identity, although other neural crest derivatives, such as melanocytes and condrocytes, have not been described [3, 8]. In our paper, we went a step further, characterizing in the differentiated cultures the full range of cranial neural crest derivatives, including melanocytes and condrocytes. In addition, we demonstrated the possibility of deriving highly proliferative neural crest precursors cell lines, a finding never reported before in previous studies on neural crest culture.

The very efficient derivation of neural crest precursors in this study is in contrast with the few reports on the characterization of neural crest derivatives in other cell systems. This fact can depend on the intrinsic dorsal identity of bovine neural rosettes, demonstrated by the expression of Pax7, and possibly on the lack of CNS inductive signals. The aim of future studies will be to further investigate this model and to clarify the effect of other inductive signals, especially those, such as retinoic acid, which have a posteriorising effect.

Another important aspect of more general interest demonstrated in this study, in a large animal model, is the possibility to derive neural rosettes and highly proliferative neural crest precursor cell lines of normal karyotype both from fertilized and from NT embryos, the latter as potential candidates for autologous cell therapy of the nuclear donor.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References
  11. Supporting Information

This study provides an original model of neural induction directly from preimplantation embryos of a large mammalian species and gives unprecedented in vitro access to early steps of nervous system development. Remarkably, this system includes both direct formation of neural tube-like structures and derivation of neural crest precursors from cloned and fertilized embryos. We show that neural rosettes are responsive to inductive signals acting in vivo and that for this reason they can represent a novel cell system on which to study early neurulation events in vitro. Finally, the characterization of highly proliferative cell lines of neural crest identity demonstrates the possibility of obtaining, directly from the embryo, peripheral nervous system and ectomesenchymal derivatives for potential cell therapy and tissue engineering applications.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References
  11. Supporting Information

We thank G. Cossu for reading of the manuscript and J.F. Brunet, G. Cossu, R. Galli, A. Gritti, V. Schiaffino, and G. Corte for sharing reagents and antisera. This work was supported by grants from Istituto Superiore di Sanità (Programma nazionale cellule staminali, CS 11 and CS 71) and FIRB (to V.B. and C.G.), Telethon (to V.B.), European Science Foundation through the CNR (Eurostells ERAS-CT.2003-980409), and Cariplo Foundation (to C.G.). Author contribution: G.L. derived the neural crest precursor cell lines and did the in vivo transplants. G.L., S.C., S.G.G., D.B., and V.B. performed the cell culture and in vitro differentiation experiments. G.L., S.C., S.G.G., and V.B. did the immunohistochemistry, and S.G.G. and V.B. did the histology. D.B. analyzed the karyotype. I.L. and C.G. provided the bovine embryos. V.B., S.G.G., and E.C. did the RT-PCR analyses. G.L., C.G., and V.B. wrote the paper.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References
  11. Supporting Information
FilenameFormatSizeDescription
Lazzari_et_al_Fig_1.6.pdf106KSupplemental Figure
Lazzari_et_al_1._Legend_of_suppl.__Fig._6_R1.pdf11KSupplemental Legend
Lazzarisuppdata.pdf11KSupplemental Data

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.