- Email: [email protected]
RT-PCR amplification of mRNA from single brain neurospheres
Journal of Neuroscience Methods 96 (2000) 57 – 61 www.elsevier.com/locate/jneumeth
RT-PCR amplification of mRNA from single brain neurospheres O.N. Suslov a, V.G. Kukekov a,b,*, E.D. Laywell a, B. Scheffler a,c, D.A. Steindler a,b a
Department of Anatomy and Neurobiology, College of Medicine, Uni6ersity of Tennessee, 855 Monroe A6enue, Memphis, TN 38163, USA b Methodist Hospitals Memphis, Brain Marrow Project, Memphis, TN, USA c Department of Neuropathology, Uni6ersity of Bonn, Bonn, Germany Received 9 April 1999; received in revised form 25 October 1999; accepted 11 November 1999
Abstract A method is described that allows cDNA production from individual brain cell clones or ‘neurospheres’. These culture-generated spheres of stem, progenitor, and differentiated cells have been the focus of interest because they represent an in vitro model of neurogenesis. However, because neurospheres are somewhat resistant, in part due to their enclosure by a dense extracellular matrix, to methods attempting to disrupt them and isolate nucleic acids, there is a need for new technology that affords the simple and efficient RT-PCR for studies of neural gene expression and discovery. A method is described here that uses sonication and an all-in-one approach for the construction of cDNA from single neurospheres. The generation of cDNA from individual adult brain stem/progenitor cell neurospheres is useful for future studies of neurogenic gene expression. © 2000 Elsevier Science B.V. All rights reserved. Keywords: mRNA amplification; Sonication; Stem/progenitor cell
1. Introduction The isolation and in vitro propagation of a putative stem cell population from the adult brain by the Weiss and Bartlett groups (Reynolds and Weiss, 1992; Richards et al., 1992) suggested that neurogenesis (the process of replication and differentiation of neural stem cells into committed progenitors and mature cells) is possible in the mature brain. Recent studies have also extended the notion of persistent neurogenesis, seen in vivo (Eriksson et al., 1998) as well as in explant cultures from the adult human hippocampus (Kirschenbaum et al., 1994), by demonstrating the existence of clonogenic stem/progenitor cells in the adult human brain that are capable of forming clones or neurospheres in vitro (Kukekov et al., 1999). ‘Neurospheres’ are thus artificial, in vitro-generated structures, where each neurosphere represents a distinct unit that arose from a single stem/progenitor cell in a particular stage of its maturation. Therefore, each neurosphere represents the clonal expansion of a cell that * Corresponding author. Tel.: +1-901-4485986; fax: + 1-901-4487193. E-mail address: [email protected] (V.G. Kukekov)
may have originated during a distinct ontological stage of neural development, and/or from a particular neurogenic region (e.g. in the adult brain, the periventricular subependymal zone that might be regionally heterogeneous, or hippocampus, see Scheffler et al., 1999 for review). It has been shown that the population of cells that constitute a neurosphere, after cultivation under particular tissue culture conditions, contains all major cell types found in the developing and adult central nervous system — neurons, astrocytes, and oligodendrocytes (Weiss et al., 1996). During their cultivation, cells in each neurosphere undergo proliferation and differentiation through a variety of stages that mirror, to an extent, cell growth and differentiation as seen in the developing brain in vivo. Thus, comparative studies of individual neurospheres could conceivably uncover stages and signals involved in the basic processes leading to neuronal or glial generation from a single stem/ progenitor cell. Each neurosphere could thus be considered an isolated developing microsystem, the study of which might provide insights into cell–cell interactions underlying neurogenesis. Furthermore, the neurosphere as an isolated developing neural microsystem can be subjected to different microenvironments (e.g. the addition of particular drugs or factors, includ-
0165-0270/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 0 2 7 0 ( 9 9 ) 0 0 1 7 7 - 6
58
O.N. Suslo6 et al. / Journal of Neuroscience Methods 96 (2000) 57–61
ing growth factors) to evaluate the consequences of varied conditions in this model neurogenic system. The present study and method is therefore based on the premise that the interclonal heterogeneity of neurospheres reflects, to some extent, the heterogeneity of stem/progenitor cell populations found in the developing and adult brain (e.g. most recently, by Kukekov et al., 1999; Palmer et al., 1999). Studies to date have focused on genetic analyses of only populations of neurospheres, perhaps due to difficulties in disrupting individual neurospheres, the shortage of material obtained from their mechanical or chemical disruption, or limited biological issues related to the isolation of genetic material from individual clones. To date, the reverse transcriptase polymerase chain reaction (RT-PCR) has been applied to populations of neurospheres for the confirmation of cell phenotype- and growth factor related molecules associated with these unique structures (Johe et al., 1996; Arsenijevic and Weiss, 1998). However, as recently emphasized, ‘‘…There are also basic technical issues relating to the growth and propagation of these cells in culture that need to be overcome. Specifically, the reported difficulty in dissociating the human ES/EG [embryonic stem and embryonic germ cell lines] cell clusters into viable single cells is problematic, particularly for genetargeting experiments…’’ (Keller and Snodgrass, 1999).
Fig. 1. Electron micrograph of an adult mouse neurosphere. Numerous tightly-packed cells are embedded within a dense matrix. There appear to be different cell types within this neurosphere, with cells having varied degrees of electron dense cytoplasm and different nuclear morphologies. Magnification bar, 10 mm. Inset shows such a neurosphere as seen and harvested through the phase microscope for cDNA generation as described here.
These stem cell-generated cell structures, like neurospheres, thus pose a similar obstacle for gene discovery studies. Neurospheres have been the focus of a great deal of attention in developmental, cellular and molecular neurobiology because they provide insights into neurogenesis, they are potentially amenable to gene discovery studies, and they also offer the possibility for future cell replacement therapies for debilitating neurological disease (Scheffler et al., 1999).
2. Materials, methods and results In this report, we describe a method that facilitates the reverse transcriptase-polymerase chain reaction (RT-PCR) amplification from single neurospheres using a sonication protocol. We developed this method because each neurosphere contains a small number of cells, approximately 50–400, which are embedded in an extremely dense extracellular matrix (see Fig. 1; and Kukekov et al., 1999) which make these structures difficult to disrupt, using conventional methods, without losing material. Therefore we have developed a fast, reliable and sensitive method for detecting gene expression in a single neurosphere. Prior to perfecting the sonication method, we also tried two methods of RNA extraction from somatic cells — the guanidine cyanide method followed by phenol-chloroform extraction, and freeze–thawing of neurospheres previously disrupted mechanically with a micromanipulator. The guanidine method of RNA preparation from single neurospheres led to a dramatic loss of material, and the freeze–thaw method was more successful but very time-consuming. There have been several methods described for the release of RNA from different sources, using sonication, but all of these have included an additional step for RNA isolation (Rajagopalan et al., 1995; Houze and Gustavsson, 1996; Liao et al., 1997) which leads to the loss of material. To avoid this, we use an approach that combines RT-PCR without RNA isolation (Fung and Fung, 1991; Kumazaki et al., 1994; O’Brien et al., 1994; von Eggeling and Ballhausen, 1995; Klebe et al., 1996) and sonication all in one tube. In the present study, we used neurospheres generated by a method that is well established in our laboratory (Kukekov et al., 1997, 1999). Single mouse neurospheres, identified under the phase or electron microscope as multicellular orbicular structures (see Fig. 1; neurosphere processing and microscopic analyses as described in Kukekov et al., 1999), were collected in a volume of 0.5 ml using a micropipette with filter tip. Each neurosphere was then transferred to a 0.6-ml tube containing 10 ml RNase-free water with 5 U RNase Inhibitor (Gibco BRL/Life Technologies, Gaithersburg, MD, USA). Neurospheres were then sonicated, using a Microtip Sonicator (Kontes, Vineland, NJ, USA), by
O.N. Suslo6 et al. / Journal of Neuroscience Methods 96 (2000) 57–61
Fig. 2. Cycle-dependent PCR product accumulation for different transcripts, testing for reproducibility of the method. Three aliquots of sonicate were subjected to cycle-dependent PCR. For each transcript, the minimum number of cycles necessary (i.e. a threshold cycle) for visualization of product is the same for all three aliquots. Lanes: 1-18S rRNA; 2-GAPDH; 3-GFAP; 4-nestin; nc-negative control.
gently touching the liquid surface for 5 s, power 4, tune 2. The tubes were kept on ice before and after sonication. The optimal time range for sonication was determined to be 4–10 s, since temperature increased during 10 s, up to 55°C. Temperature was measured in the test tubes, during sonication, using a digital mini-thermometer HH81 (Omega Engineering Inc., Stamford, CT, USA). It is not recommend to sonicate less than 4 s, to assure that RNA is completely released, but no more than 10 s, since raising the temperature decreases RNase inhibitor activity. While working with a number of samples, to avoid overheating of the microtip and cross contamination of samples, the sonicator microtip should be rinsed in consecutive solutions of ice cold 1 M HCl, 1 M NaOH, 1 M Tris – HCl, pH 7.5, doubledistilled H2O. First-strand cDNA was prepared with Superscript reverse transcriptase (Gibco BRL). Briefly, the tube with a single sonicated neurosphere and 0.5 ml oligo(dT)12 – 18 was heated to 70°C for 2 min and then chilled on ice. The annealed mRNA-oligo(dT)12 – 18 complexes were incubated at 42°C for 60 min with 200 U of reverse transcriptase, 20 mM Tris, pH 8.4, 50 mM KCl, 2.5 mM MgCl2, 10 mM DTT, 0.5 mM dNTPs, 5 U RNase Inhibitor in 20 ml total volume. The RNA template was then removed from the cDNA:RNA hybrid by the addition of 4 U RNase H (37°C, 20 min). The 1 ml of reaction product was added to 29 ml of the PCR mixture (0.67 mM of each primer and 1 U Platinum Taq DNA polymerase (Gibco BRL), 0.2 mM dNTPs, 1X PCR buffer (Gibco BRL), and 1.5 mM MgCl2). The conditions for the PCR reaction were optimized for each set of primers, and PCR analyses were performed using touchdown (TD)-PCR, on a MJR PCR machine. PCR products were analyzed by electrophoresis in a 1.7% agarose gel containing ethidium bromide for visualization.
59
For the evaluation of the method, we used transcripts of three genes known to be expressed by neurosphere cells (Reynolds and Weiss, 1992; Richards et al., 1992; Weiss et al., 1996; Johe et al., 1996; Arsenijevic and Weiss, 1998) — glyceraldehyde-3phosphate dehydrogenase (GAPDH), glial fibrillary acidic protein (GFAP; the intermediate filament protein found in astrocytes) and nestin (the intermediate filament protein found in brain stem and precursor cell populations). 18S rRNA was used as a control because it’s level is proportional to the number of cells (i.e. the content of ribosomal RNA in each neurosphere cell is presumed to be approximately the same). Primers were designed with the program Oligo 5.0 from published sequence data as follows: (1) 18S rRNA, the upstream primer — 5%-CTGAGAAACGGCTACCACATCC-3% and the downstream primer — 5%-CGCGGTCCTATTCCATTATTCCT-3%, 439 bp amplified product, (2) GAPDH, the upstream primer — 5%-GTCGGTGTGAACGGATTTG-3% and the downstream primer — 5%-TAGACTCCACGACATACTCAGCA-3%, 280 bp amplified product, (3) GFAP, the upstream primer — 5%-GCCAAGGAGCCCACCAAACT-3% and the downstream primer — 5%-ATCTCCTCCTCCAGCGATTCAAC-3%, 275 bp amplified product, (4) nestin, the upstream primer — 5%-GAAGCCCTGGAGCAGGAGAAGCA-3% and the downstream primer — 5%TCCAGGTGTCTGCAACCGAGAGTTC-3%, 159 bp amplified product. All primers were designed to span intron/exon junction(s). The upper primers for GAPDH, GFAP and nestin were ordered for the first coding exon. To evaluate the reproducibility of results, three different neurospheres were combined in one tube, sonicated, and three aliquots were split for further RT-PCR analysis. Cycle-dependent PCR of each aliquot, with pairs of primers for 18S rRNA, GAPDH, GFAP, and nestin was performed for RT product. During PCR after a certain number of cycles, samples of product were taken and loaded to 1.7% agarose gels (Fig. 2). The results reveal a sufficient reproducibility for all aliquots looking at four different templates. The distribution of mRNA recovery across neurospheres of different sizes also was studied. Neurosphere size has been shown to relate to different types of neurospheres, including a range from the most immature (smallest) to the most differentiated (largest) (Kukekov et al., 1997, 1999). These experiments were also conducted in order to evaluate the cycle-dependent accumulation of PCR products for each primer pair. Three different size neurospheres (large, medium, and small) were sonicated in separate tubes, and the sonicate was used for RT-PCR according to the protocol described above. Cycle-dependent PCR (see above), with pairs of primers for 18S rRNA and GAPDH, was performed. The results obtained for both transcripts
60
O.N. Suslo6 et al. / Journal of Neuroscience Methods 96 (2000) 57–61
revealed that the number of cells in a neurosphere correlates with neurosphere size (Fig. 3). An experiment also was conducted to check crosscontamination (Fig. 4). We used single mouse and human neurospheres in the same experiment. Single human and mouse neurospheres were sonicated, and sonicates were used for RT-PCR with pairs of primers for unique regions of human and mouse GAPDH. Pairs of primers for GAPDH were used, for a unique part of the gene, as follows: (1) GAPDH, mouse, the upstream primer — 5%-TTCCTACCCCCAATGTGTCCGTC-3% and the downstream primer — 5%-ACCCTGTTGCTGTAGCCGTATTCA-3%, 268 bp amplified product, (2) GAPDH, human, the upstream primer — 5%-TCCCCACTGCCAACGTGTCAGTG-3% and the downstream primer — 5%-ACCCTGTTGCTGTAGCCAAATTCG-3%, 268 bp amplified product. After 50 cycles of PCR, we were not able to detect any visible cross contamination in the test tubes. Finally, in order to check the integrity of different templates, we utilized upper primers for the first coding exon. Taking into account that reverse transcription was made using oligo(dT)12 – l8 primers, we propose that
Fig. 3. Cycle-dependent PCR product accumulation for different size neurospheres to demonstrate the distribution of mRNA recovery across neurospheres. This figure indicates that mRNA recovery is proportional to the number of cells within single neurospheres (i.e. correlated with neurosphere size). Lane 1-18S rRNA; Lane 2 GAPDH.
Fig. 4. A check for cross-contamination using single mouse and human neurospheres in the same experiment. The described method of sonication and RT-PCR was applied to single neurospheres from these different species, processed identically and at the same time. Using primers for human and mouse GAPDH, 50 cycles of Touchdown PCR were performed for each sample, and it is shown that there is an absence of human transcripts in the mouse sample, and likewise, no mouse transcripts in the human sample. L-5 0 bp ladder, (Gibco BRL); M-mouse sphere; H-human sphere; nc -negative control.
we are dealing with at least full length coding structure plus the 3%-untranslated region (GAPDH-1.2 kB, GFAP-2.5 kB, nestin-5.5 kB). The products of singlesphere RT-PCR were sequenced (Automatic sequencer, St. Jude Children’s Research Hospital).
3. Discussion The method we describe here provides the opportunity to perform RT-PCR assays without time-consuming RNA isolation. All manipulations are carried out in one tube, and amplified cDNA is produced from all RNA available from small numbers of cells. This method is thus useful in studies of brain neurospheres where amplification of RNA can be used to study developmental gene expression. A method for the rapid and efficient amplification of RNA from neurospheres derived from valuable tissue specimens is, therefore, useful for the identification of genes that might be involved in normal brain development, as well as in neurological disease including neurodegeneration and brain neoplasia. This method is reproducible, avoids cross-contamination, and provides a reliable recovery and preserved integrity of mRNA from a heterogeneous population of neurospheres. The efficiency and sensitivity of the method described here has widespread applicability. It can be used to amplify RNA from difficult to disrupt cellular structures like neurospheres [which are tight clones of cells surrounded by extracellular matrix molecules (Kukekov et al., 1999) that make them rather resistant using mechanical or other methods of disruption]. This study is the first to create cDNA uncloned libraries from individual neurospheres. Previous studies (Johe et al., 1996; Arsenijevic and Weiss, 1998) have used RT-PCR for populations of neurospheres to show the presence of transcripts for well-known neural cell phenotype genes and growth factor-related molecules that are crucial to the clonal expansion and differentiation of stem/progenitor cells. The method described here likewise has the potential for future use in studies attempting to compare individual neurosphere gene expression as a model system for understanding neurogenic gene expression. That is, in light of individual neurospheres varying in cellular and molecular composition (Kukekov et al., 1997; Svendsen et al., 1998; Scheffler et al., 1999), presumably as a result of their generation from different stem/progenitor cell types and/or variation due to their in vitro isolation from different regions or at different stages of their evolution and maturation, they offer a unique model system for determining the sequential patterns of gene expression during in vitro neurogenesis. Such a model system, that offers the potential for establishing the precise temporal
O.N. Suslo6 et al. / Journal of Neuroscience Methods 96 (2000) 57–61
sequence of gene expression in a close population of tightly controlled stem/progenitor and differentiating neural cell clones, also offers the potential for novel gene discovery. The neurosphere is thus a potentially useful model system for studies of the cell and molecular biology of neurogenesis, and the technique described here provides a sensitive and efficient method for amplifying message from such an individual neural cell clone. Future studies might rely on methods such as this, in combination with subtractive hybridization comparing individual neurospheres, to produce libraries of differentially expressed transcripts. Such libraries could also be used to generate microarrays and gene chips for the characterization of new as well as existing genes involved in stem/progenitor cell propagation and differentiation.
Acknowledgements This work was supported by NIH/NINDS grants NS29225, NS37556, a grant from the Spinal Cord Research Foundation, Paralyzed Veterans of America, and the Methodist LeBonheur Healthcare Foundation. The authors would like to thank Armin Buss for technical assistance, and Dennis W. Martin for advice with the temperature measurements.
References Arsenijevic Y, Weiss S. Insulin-like growth factor-I is a differentiation factor for postmitotic CNS stem cell-derived neuronal precursors: distinct actions from those of brain-derived neurotrophic factor. J Neurosci 1998;18:2118–28. Eriksson PS, Perfilieva E, Bjork-Eriksson T, Alborn A-M, Nordborg C, Peterson DA, Gage FH. Neurogenesis in the adult human hippocampus. Nature Med 1998;4:1313–7. Fung MC, Fung KY. PCR amplification of mRNA directly from a crude cell lysate prepared by thermophilic protease digestion. Nucleic Acids Res 1991;19:4300. Houze TA, Gustavsson B. Sonification as a means of enhancing the detection of gene expression levels from formalin-fixed, paraffinembedded biopsies. Biotechniques 1996;21:1074–8. Johe KK, Hazel TG, Muller T, Dugich-Djordjevic MM, McKay RDG. Single factors direct the differentiation of stem cells from
.
61
the fetal and adult central nervous system. Genes Dev 1996;10:3129 – 40. Keller G, Snodgrass HR. Human embryonic stem cells: the future is now. Nature Med 1999;5:151 – 2. Kirschenbaum B, Nedergarrd M, Preuss A, Barami K, Fraser R, Goldman S. In vitro production and differentiation by precursor cells derived from the adult human forebrain. Cereb Cortex 1994;6:576 – 89. Klebe RJ, Grant GM, Grant AM, Garcia MA, Giambernardi TA, Taylor GP. RT-PCR without RNA isolation. Biotechniques 1996;21:1094 – 100. Kukekov VG, Laywell ED, Thomas LB, Steindler DA. A nestin-negative precursor cell from the adult mouse brain gives rise to neurons and glia. GLIA 1997;21:399 – 407. Kukekov VG, Laywell ED, Suslov ON, Thomas LB, Scheffler B, Davies K, O’Brien TF, Kusakabe M, Steindler DA. Multipotent stem/progenitor cells with similar properties arise from two neurogenic regions of adult human brain. Exp Neurol 1999;156:333–44. Kumazaki T, Hamada K, Mitsui Y. Detection of mRNA expression in a single cell by direct RT-PCR. Biotechniques 1994;16:1017–9. Liao YC, Drossard J, Nahring JM, Fischer R. Isolation of RNA from plant cell suspension cultures and calli by sonication. Biotechniques 1997;23:996 – 8. O’Brien DP, Billadeau D, Van Ness B. RT-PCR assay for detection of transcripts from very few cells using whole cell lysates. Biotechniques 1994;16:586 – 8. Palmer TD, Markakis EA, Willhoite AR, Safar F, Gage FH. Fibroblast growth factor-2 activates a latent neurogenic program in neural stem cells from diverse regions of the adult CNS. J Neurosci 1999;19:8487 – 97. Rajagopalan M, Boggaram V, Madiraju MV. A rapid protocol for isolation of RNA from mycobacteria. Lett Appl Microbiol 1995;21:14 – 7. Reynolds BA, Weiss S. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 1992;255:1707– 10. Richards LJ, Kilpatrick TJ, Bartlett PF. De novo generation of neuronal cells from the adult mouse brain. Proc Natl Acad Sci 1992;89:8591 – 5. Scheffler B, Horn M, Bluemcke I, Kukekov V, Laywell ED, Steindler DA. Marrow-mindedness: a perspective on neuropoiesis. Trends Neurosci 1999;22:348 – 57. Svendsen CN, ter Borg MG, Armstrong RJE, Rosser AE, Chandran S, Otenfeld T, Caldwell MA. A new method for the rapid and long term growth of human neural precursor cells. J Neurosci Methods 1998;85:41 – 52. von Eggeling F, Ballhausen W. Freezing of isolated cells provides free mRNA for RT-PCR amplification. Biotechniques 1995;18:408–9. Weiss S, Dunne C, Hewson J, Wohl C, Wheatly M, Peterson AC, Reynolds BA. Multipotent CNS stem cells are present in the adult mammalian spinal cord and ventricular neuroaxis. J Neurosci 1996;16:7599 – 609.
RT-PCR amplification of mRNA from single brain neurospheres O.N. Suslov a, V.G. Kukekov a,b,*, E.D. Laywell a, B. Scheffler a,c, D.A. Steindler a,b a
Department of Anatomy and Neurobiology, College of Medicine, Uni6ersity of Tennessee, 855 Monroe A6enue, Memphis, TN 38163, USA b Methodist Hospitals Memphis, Brain Marrow Project, Memphis, TN, USA c Department of Neuropathology, Uni6ersity of Bonn, Bonn, Germany Received 9 April 1999; received in revised form 25 October 1999; accepted 11 November 1999
Abstract A method is described that allows cDNA production from individual brain cell clones or ‘neurospheres’. These culture-generated spheres of stem, progenitor, and differentiated cells have been the focus of interest because they represent an in vitro model of neurogenesis. However, because neurospheres are somewhat resistant, in part due to their enclosure by a dense extracellular matrix, to methods attempting to disrupt them and isolate nucleic acids, there is a need for new technology that affords the simple and efficient RT-PCR for studies of neural gene expression and discovery. A method is described here that uses sonication and an all-in-one approach for the construction of cDNA from single neurospheres. The generation of cDNA from individual adult brain stem/progenitor cell neurospheres is useful for future studies of neurogenic gene expression. © 2000 Elsevier Science B.V. All rights reserved. Keywords: mRNA amplification; Sonication; Stem/progenitor cell
1. Introduction The isolation and in vitro propagation of a putative stem cell population from the adult brain by the Weiss and Bartlett groups (Reynolds and Weiss, 1992; Richards et al., 1992) suggested that neurogenesis (the process of replication and differentiation of neural stem cells into committed progenitors and mature cells) is possible in the mature brain. Recent studies have also extended the notion of persistent neurogenesis, seen in vivo (Eriksson et al., 1998) as well as in explant cultures from the adult human hippocampus (Kirschenbaum et al., 1994), by demonstrating the existence of clonogenic stem/progenitor cells in the adult human brain that are capable of forming clones or neurospheres in vitro (Kukekov et al., 1999). ‘Neurospheres’ are thus artificial, in vitro-generated structures, where each neurosphere represents a distinct unit that arose from a single stem/progenitor cell in a particular stage of its maturation. Therefore, each neurosphere represents the clonal expansion of a cell that * Corresponding author. Tel.: +1-901-4485986; fax: + 1-901-4487193. E-mail address: [email protected] (V.G. Kukekov)
may have originated during a distinct ontological stage of neural development, and/or from a particular neurogenic region (e.g. in the adult brain, the periventricular subependymal zone that might be regionally heterogeneous, or hippocampus, see Scheffler et al., 1999 for review). It has been shown that the population of cells that constitute a neurosphere, after cultivation under particular tissue culture conditions, contains all major cell types found in the developing and adult central nervous system — neurons, astrocytes, and oligodendrocytes (Weiss et al., 1996). During their cultivation, cells in each neurosphere undergo proliferation and differentiation through a variety of stages that mirror, to an extent, cell growth and differentiation as seen in the developing brain in vivo. Thus, comparative studies of individual neurospheres could conceivably uncover stages and signals involved in the basic processes leading to neuronal or glial generation from a single stem/ progenitor cell. Each neurosphere could thus be considered an isolated developing microsystem, the study of which might provide insights into cell–cell interactions underlying neurogenesis. Furthermore, the neurosphere as an isolated developing neural microsystem can be subjected to different microenvironments (e.g. the addition of particular drugs or factors, includ-
0165-0270/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 5 - 0 2 7 0 ( 9 9 ) 0 0 1 7 7 - 6
58
O.N. Suslo6 et al. / Journal of Neuroscience Methods 96 (2000) 57–61
ing growth factors) to evaluate the consequences of varied conditions in this model neurogenic system. The present study and method is therefore based on the premise that the interclonal heterogeneity of neurospheres reflects, to some extent, the heterogeneity of stem/progenitor cell populations found in the developing and adult brain (e.g. most recently, by Kukekov et al., 1999; Palmer et al., 1999). Studies to date have focused on genetic analyses of only populations of neurospheres, perhaps due to difficulties in disrupting individual neurospheres, the shortage of material obtained from their mechanical or chemical disruption, or limited biological issues related to the isolation of genetic material from individual clones. To date, the reverse transcriptase polymerase chain reaction (RT-PCR) has been applied to populations of neurospheres for the confirmation of cell phenotype- and growth factor related molecules associated with these unique structures (Johe et al., 1996; Arsenijevic and Weiss, 1998). However, as recently emphasized, ‘‘…There are also basic technical issues relating to the growth and propagation of these cells in culture that need to be overcome. Specifically, the reported difficulty in dissociating the human ES/EG [embryonic stem and embryonic germ cell lines] cell clusters into viable single cells is problematic, particularly for genetargeting experiments…’’ (Keller and Snodgrass, 1999).
Fig. 1. Electron micrograph of an adult mouse neurosphere. Numerous tightly-packed cells are embedded within a dense matrix. There appear to be different cell types within this neurosphere, with cells having varied degrees of electron dense cytoplasm and different nuclear morphologies. Magnification bar, 10 mm. Inset shows such a neurosphere as seen and harvested through the phase microscope for cDNA generation as described here.
These stem cell-generated cell structures, like neurospheres, thus pose a similar obstacle for gene discovery studies. Neurospheres have been the focus of a great deal of attention in developmental, cellular and molecular neurobiology because they provide insights into neurogenesis, they are potentially amenable to gene discovery studies, and they also offer the possibility for future cell replacement therapies for debilitating neurological disease (Scheffler et al., 1999).
2. Materials, methods and results In this report, we describe a method that facilitates the reverse transcriptase-polymerase chain reaction (RT-PCR) amplification from single neurospheres using a sonication protocol. We developed this method because each neurosphere contains a small number of cells, approximately 50–400, which are embedded in an extremely dense extracellular matrix (see Fig. 1; and Kukekov et al., 1999) which make these structures difficult to disrupt, using conventional methods, without losing material. Therefore we have developed a fast, reliable and sensitive method for detecting gene expression in a single neurosphere. Prior to perfecting the sonication method, we also tried two methods of RNA extraction from somatic cells — the guanidine cyanide method followed by phenol-chloroform extraction, and freeze–thawing of neurospheres previously disrupted mechanically with a micromanipulator. The guanidine method of RNA preparation from single neurospheres led to a dramatic loss of material, and the freeze–thaw method was more successful but very time-consuming. There have been several methods described for the release of RNA from different sources, using sonication, but all of these have included an additional step for RNA isolation (Rajagopalan et al., 1995; Houze and Gustavsson, 1996; Liao et al., 1997) which leads to the loss of material. To avoid this, we use an approach that combines RT-PCR without RNA isolation (Fung and Fung, 1991; Kumazaki et al., 1994; O’Brien et al., 1994; von Eggeling and Ballhausen, 1995; Klebe et al., 1996) and sonication all in one tube. In the present study, we used neurospheres generated by a method that is well established in our laboratory (Kukekov et al., 1997, 1999). Single mouse neurospheres, identified under the phase or electron microscope as multicellular orbicular structures (see Fig. 1; neurosphere processing and microscopic analyses as described in Kukekov et al., 1999), were collected in a volume of 0.5 ml using a micropipette with filter tip. Each neurosphere was then transferred to a 0.6-ml tube containing 10 ml RNase-free water with 5 U RNase Inhibitor (Gibco BRL/Life Technologies, Gaithersburg, MD, USA). Neurospheres were then sonicated, using a Microtip Sonicator (Kontes, Vineland, NJ, USA), by
O.N. Suslo6 et al. / Journal of Neuroscience Methods 96 (2000) 57–61
Fig. 2. Cycle-dependent PCR product accumulation for different transcripts, testing for reproducibility of the method. Three aliquots of sonicate were subjected to cycle-dependent PCR. For each transcript, the minimum number of cycles necessary (i.e. a threshold cycle) for visualization of product is the same for all three aliquots. Lanes: 1-18S rRNA; 2-GAPDH; 3-GFAP; 4-nestin; nc-negative control.
gently touching the liquid surface for 5 s, power 4, tune 2. The tubes were kept on ice before and after sonication. The optimal time range for sonication was determined to be 4–10 s, since temperature increased during 10 s, up to 55°C. Temperature was measured in the test tubes, during sonication, using a digital mini-thermometer HH81 (Omega Engineering Inc., Stamford, CT, USA). It is not recommend to sonicate less than 4 s, to assure that RNA is completely released, but no more than 10 s, since raising the temperature decreases RNase inhibitor activity. While working with a number of samples, to avoid overheating of the microtip and cross contamination of samples, the sonicator microtip should be rinsed in consecutive solutions of ice cold 1 M HCl, 1 M NaOH, 1 M Tris – HCl, pH 7.5, doubledistilled H2O. First-strand cDNA was prepared with Superscript reverse transcriptase (Gibco BRL). Briefly, the tube with a single sonicated neurosphere and 0.5 ml oligo(dT)12 – 18 was heated to 70°C for 2 min and then chilled on ice. The annealed mRNA-oligo(dT)12 – 18 complexes were incubated at 42°C for 60 min with 200 U of reverse transcriptase, 20 mM Tris, pH 8.4, 50 mM KCl, 2.5 mM MgCl2, 10 mM DTT, 0.5 mM dNTPs, 5 U RNase Inhibitor in 20 ml total volume. The RNA template was then removed from the cDNA:RNA hybrid by the addition of 4 U RNase H (37°C, 20 min). The 1 ml of reaction product was added to 29 ml of the PCR mixture (0.67 mM of each primer and 1 U Platinum Taq DNA polymerase (Gibco BRL), 0.2 mM dNTPs, 1X PCR buffer (Gibco BRL), and 1.5 mM MgCl2). The conditions for the PCR reaction were optimized for each set of primers, and PCR analyses were performed using touchdown (TD)-PCR, on a MJR PCR machine. PCR products were analyzed by electrophoresis in a 1.7% agarose gel containing ethidium bromide for visualization.
59
For the evaluation of the method, we used transcripts of three genes known to be expressed by neurosphere cells (Reynolds and Weiss, 1992; Richards et al., 1992; Weiss et al., 1996; Johe et al., 1996; Arsenijevic and Weiss, 1998) — glyceraldehyde-3phosphate dehydrogenase (GAPDH), glial fibrillary acidic protein (GFAP; the intermediate filament protein found in astrocytes) and nestin (the intermediate filament protein found in brain stem and precursor cell populations). 18S rRNA was used as a control because it’s level is proportional to the number of cells (i.e. the content of ribosomal RNA in each neurosphere cell is presumed to be approximately the same). Primers were designed with the program Oligo 5.0 from published sequence data as follows: (1) 18S rRNA, the upstream primer — 5%-CTGAGAAACGGCTACCACATCC-3% and the downstream primer — 5%-CGCGGTCCTATTCCATTATTCCT-3%, 439 bp amplified product, (2) GAPDH, the upstream primer — 5%-GTCGGTGTGAACGGATTTG-3% and the downstream primer — 5%-TAGACTCCACGACATACTCAGCA-3%, 280 bp amplified product, (3) GFAP, the upstream primer — 5%-GCCAAGGAGCCCACCAAACT-3% and the downstream primer — 5%-ATCTCCTCCTCCAGCGATTCAAC-3%, 275 bp amplified product, (4) nestin, the upstream primer — 5%-GAAGCCCTGGAGCAGGAGAAGCA-3% and the downstream primer — 5%TCCAGGTGTCTGCAACCGAGAGTTC-3%, 159 bp amplified product. All primers were designed to span intron/exon junction(s). The upper primers for GAPDH, GFAP and nestin were ordered for the first coding exon. To evaluate the reproducibility of results, three different neurospheres were combined in one tube, sonicated, and three aliquots were split for further RT-PCR analysis. Cycle-dependent PCR of each aliquot, with pairs of primers for 18S rRNA, GAPDH, GFAP, and nestin was performed for RT product. During PCR after a certain number of cycles, samples of product were taken and loaded to 1.7% agarose gels (Fig. 2). The results reveal a sufficient reproducibility for all aliquots looking at four different templates. The distribution of mRNA recovery across neurospheres of different sizes also was studied. Neurosphere size has been shown to relate to different types of neurospheres, including a range from the most immature (smallest) to the most differentiated (largest) (Kukekov et al., 1997, 1999). These experiments were also conducted in order to evaluate the cycle-dependent accumulation of PCR products for each primer pair. Three different size neurospheres (large, medium, and small) were sonicated in separate tubes, and the sonicate was used for RT-PCR according to the protocol described above. Cycle-dependent PCR (see above), with pairs of primers for 18S rRNA and GAPDH, was performed. The results obtained for both transcripts
60
O.N. Suslo6 et al. / Journal of Neuroscience Methods 96 (2000) 57–61
revealed that the number of cells in a neurosphere correlates with neurosphere size (Fig. 3). An experiment also was conducted to check crosscontamination (Fig. 4). We used single mouse and human neurospheres in the same experiment. Single human and mouse neurospheres were sonicated, and sonicates were used for RT-PCR with pairs of primers for unique regions of human and mouse GAPDH. Pairs of primers for GAPDH were used, for a unique part of the gene, as follows: (1) GAPDH, mouse, the upstream primer — 5%-TTCCTACCCCCAATGTGTCCGTC-3% and the downstream primer — 5%-ACCCTGTTGCTGTAGCCGTATTCA-3%, 268 bp amplified product, (2) GAPDH, human, the upstream primer — 5%-TCCCCACTGCCAACGTGTCAGTG-3% and the downstream primer — 5%-ACCCTGTTGCTGTAGCCAAATTCG-3%, 268 bp amplified product. After 50 cycles of PCR, we were not able to detect any visible cross contamination in the test tubes. Finally, in order to check the integrity of different templates, we utilized upper primers for the first coding exon. Taking into account that reverse transcription was made using oligo(dT)12 – l8 primers, we propose that
Fig. 3. Cycle-dependent PCR product accumulation for different size neurospheres to demonstrate the distribution of mRNA recovery across neurospheres. This figure indicates that mRNA recovery is proportional to the number of cells within single neurospheres (i.e. correlated with neurosphere size). Lane 1-18S rRNA; Lane 2 GAPDH.
Fig. 4. A check for cross-contamination using single mouse and human neurospheres in the same experiment. The described method of sonication and RT-PCR was applied to single neurospheres from these different species, processed identically and at the same time. Using primers for human and mouse GAPDH, 50 cycles of Touchdown PCR were performed for each sample, and it is shown that there is an absence of human transcripts in the mouse sample, and likewise, no mouse transcripts in the human sample. L-5 0 bp ladder, (Gibco BRL); M-mouse sphere; H-human sphere; nc -negative control.
we are dealing with at least full length coding structure plus the 3%-untranslated region (GAPDH-1.2 kB, GFAP-2.5 kB, nestin-5.5 kB). The products of singlesphere RT-PCR were sequenced (Automatic sequencer, St. Jude Children’s Research Hospital).
3. Discussion The method we describe here provides the opportunity to perform RT-PCR assays without time-consuming RNA isolation. All manipulations are carried out in one tube, and amplified cDNA is produced from all RNA available from small numbers of cells. This method is thus useful in studies of brain neurospheres where amplification of RNA can be used to study developmental gene expression. A method for the rapid and efficient amplification of RNA from neurospheres derived from valuable tissue specimens is, therefore, useful for the identification of genes that might be involved in normal brain development, as well as in neurological disease including neurodegeneration and brain neoplasia. This method is reproducible, avoids cross-contamination, and provides a reliable recovery and preserved integrity of mRNA from a heterogeneous population of neurospheres. The efficiency and sensitivity of the method described here has widespread applicability. It can be used to amplify RNA from difficult to disrupt cellular structures like neurospheres [which are tight clones of cells surrounded by extracellular matrix molecules (Kukekov et al., 1999) that make them rather resistant using mechanical or other methods of disruption]. This study is the first to create cDNA uncloned libraries from individual neurospheres. Previous studies (Johe et al., 1996; Arsenijevic and Weiss, 1998) have used RT-PCR for populations of neurospheres to show the presence of transcripts for well-known neural cell phenotype genes and growth factor-related molecules that are crucial to the clonal expansion and differentiation of stem/progenitor cells. The method described here likewise has the potential for future use in studies attempting to compare individual neurosphere gene expression as a model system for understanding neurogenic gene expression. That is, in light of individual neurospheres varying in cellular and molecular composition (Kukekov et al., 1997; Svendsen et al., 1998; Scheffler et al., 1999), presumably as a result of their generation from different stem/progenitor cell types and/or variation due to their in vitro isolation from different regions or at different stages of their evolution and maturation, they offer a unique model system for determining the sequential patterns of gene expression during in vitro neurogenesis. Such a model system, that offers the potential for establishing the precise temporal
O.N. Suslo6 et al. / Journal of Neuroscience Methods 96 (2000) 57–61
sequence of gene expression in a close population of tightly controlled stem/progenitor and differentiating neural cell clones, also offers the potential for novel gene discovery. The neurosphere is thus a potentially useful model system for studies of the cell and molecular biology of neurogenesis, and the technique described here provides a sensitive and efficient method for amplifying message from such an individual neural cell clone. Future studies might rely on methods such as this, in combination with subtractive hybridization comparing individual neurospheres, to produce libraries of differentially expressed transcripts. Such libraries could also be used to generate microarrays and gene chips for the characterization of new as well as existing genes involved in stem/progenitor cell propagation and differentiation.
Acknowledgements This work was supported by NIH/NINDS grants NS29225, NS37556, a grant from the Spinal Cord Research Foundation, Paralyzed Veterans of America, and the Methodist LeBonheur Healthcare Foundation. The authors would like to thank Armin Buss for technical assistance, and Dennis W. Martin for advice with the temperature measurements.
References Arsenijevic Y, Weiss S. Insulin-like growth factor-I is a differentiation factor for postmitotic CNS stem cell-derived neuronal precursors: distinct actions from those of brain-derived neurotrophic factor. J Neurosci 1998;18:2118–28. Eriksson PS, Perfilieva E, Bjork-Eriksson T, Alborn A-M, Nordborg C, Peterson DA, Gage FH. Neurogenesis in the adult human hippocampus. Nature Med 1998;4:1313–7. Fung MC, Fung KY. PCR amplification of mRNA directly from a crude cell lysate prepared by thermophilic protease digestion. Nucleic Acids Res 1991;19:4300. Houze TA, Gustavsson B. Sonification as a means of enhancing the detection of gene expression levels from formalin-fixed, paraffinembedded biopsies. Biotechniques 1996;21:1074–8. Johe KK, Hazel TG, Muller T, Dugich-Djordjevic MM, McKay RDG. Single factors direct the differentiation of stem cells from
.
61
the fetal and adult central nervous system. Genes Dev 1996;10:3129 – 40. Keller G, Snodgrass HR. Human embryonic stem cells: the future is now. Nature Med 1999;5:151 – 2. Kirschenbaum B, Nedergarrd M, Preuss A, Barami K, Fraser R, Goldman S. In vitro production and differentiation by precursor cells derived from the adult human forebrain. Cereb Cortex 1994;6:576 – 89. Klebe RJ, Grant GM, Grant AM, Garcia MA, Giambernardi TA, Taylor GP. RT-PCR without RNA isolation. Biotechniques 1996;21:1094 – 100. Kukekov VG, Laywell ED, Thomas LB, Steindler DA. A nestin-negative precursor cell from the adult mouse brain gives rise to neurons and glia. GLIA 1997;21:399 – 407. Kukekov VG, Laywell ED, Suslov ON, Thomas LB, Scheffler B, Davies K, O’Brien TF, Kusakabe M, Steindler DA. Multipotent stem/progenitor cells with similar properties arise from two neurogenic regions of adult human brain. Exp Neurol 1999;156:333–44. Kumazaki T, Hamada K, Mitsui Y. Detection of mRNA expression in a single cell by direct RT-PCR. Biotechniques 1994;16:1017–9. Liao YC, Drossard J, Nahring JM, Fischer R. Isolation of RNA from plant cell suspension cultures and calli by sonication. Biotechniques 1997;23:996 – 8. O’Brien DP, Billadeau D, Van Ness B. RT-PCR assay for detection of transcripts from very few cells using whole cell lysates. Biotechniques 1994;16:586 – 8. Palmer TD, Markakis EA, Willhoite AR, Safar F, Gage FH. Fibroblast growth factor-2 activates a latent neurogenic program in neural stem cells from diverse regions of the adult CNS. J Neurosci 1999;19:8487 – 97. Rajagopalan M, Boggaram V, Madiraju MV. A rapid protocol for isolation of RNA from mycobacteria. Lett Appl Microbiol 1995;21:14 – 7. Reynolds BA, Weiss S. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 1992;255:1707– 10. Richards LJ, Kilpatrick TJ, Bartlett PF. De novo generation of neuronal cells from the adult mouse brain. Proc Natl Acad Sci 1992;89:8591 – 5. Scheffler B, Horn M, Bluemcke I, Kukekov V, Laywell ED, Steindler DA. Marrow-mindedness: a perspective on neuropoiesis. Trends Neurosci 1999;22:348 – 57. Svendsen CN, ter Borg MG, Armstrong RJE, Rosser AE, Chandran S, Otenfeld T, Caldwell MA. A new method for the rapid and long term growth of human neural precursor cells. J Neurosci Methods 1998;85:41 – 52. von Eggeling F, Ballhausen W. Freezing of isolated cells provides free mRNA for RT-PCR amplification. Biotechniques 1995;18:408–9. Weiss S, Dunne C, Hewson J, Wohl C, Wheatly M, Peterson AC, Reynolds BA. Multipotent CNS stem cells are present in the adult mammalian spinal cord and ventricular neuroaxis. J Neurosci 1996;16:7599 – 609.