Engineering embryonic stem cell derived glia for adenosine delivery

Neuroscience Letters 370 (2004) 160–165

Engineering embryonic stem cell derived glia for adenosine delivery Denise E. Fedelea , Peter Kochb , Louis Scheurera , Elizabeth M. Simpsonc , Hanns M¨ohlera,d , Oliver Br¨ustleb , Detlev Boisona,d,∗ a

Institute of Pharmacology and Toxicology, University of Z¨urich, CH-8057 Z¨urich, Switzerland Institute of Reconstructive Neurobiology, University of Bonn Medical Center and Hertie Foundation, D-53105 Bonn, Germany Centre for Molecular Medicine and Therapeutics, British Columbia Research Institute for Children’s and Women’s Health and Department of Medical Genetics, University of British Columbia, Vancouver, Canada V5Z 4H4 d Institute of Pharmaceutical Sciences, Federal Institute of Technology (ETH), CH-8057 Z¨ urich, Switzerland b

c

Received 15 July 2004; received in revised form 9 August 2004; accepted 9 August 2004

Abstract Based on the anticonvulsant and neuroprotective properties of adenosine, and based on the long-term survival potential of stem cell derived brain implants, adenosine releasing stem cells may constitute a novel tool for the treatment of epilepsy. Pluripotency and unlimited self-renewal make embryonic stem (ES) cells a particularly versatile donor source for cell transplantation. With the aim to test the feasibility of a stem cell-based delivery system for adenosine, both alleles of adenosine kinase (ADK), the major adenosine-metabolizing enzyme, were disrupted by homologous recombination in ES cells. Adk−/− ES cells were subjected to a glial differentiation protocol and, as a result, gave rise to proliferating glial precursors, which could be further differentiated into mature astrocytes and oligodendrocytes. Thus, a lack of ADK does not compromise the glial differentiation potential of ES cells. The Adk−/− ES cells yielded glial populations with an adenosine release of up to 40.1 ± 6.0 ng per 105 cells per hour, an amount considered to be sufficient for seizure suppression. Our findings indicate that Adk−/− ES cells constitute a potential source for therapeutic adenosine releasing grafts. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Adenosine; Adenosine kinase; ES cells; Epilepsy; Gene therapy; Glial differentiation

Adenosine and its A1 receptor agonists have potent inhibitory effects on neuronal activity and are effective in seizure suppression and neuroprotection [2,7,15]. The recent finding that activation of A1 receptors suppresses seizures in a model of pharmacoresistant epilepsy depicts adenosine as a potential therapeutic tool surpassing the efficiency of conventional antiepileptic drugs [9]. However, due to peripheral actions of adenosine, its systemic application is precluded, thus requiring a local mode of application, which could be achieved by intracerebral cellular adenosine releasing implants. The accumulation and release of adenosine is mainly controlled by the activity of adenosine kinase (ADK; EC 2.7.1.20), the key enzyme of adenosine metabolism. Consequently, inhibition of ADK leads to rapid and large increases in adenosine. Con∗

Corresponding author. Tel.: +41 1 635 59 31; fax: +41 1 635 68 74. E-mail address: [email protected] (D. Boison).

0304-3940/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2004.08.031

versely, overexpression of ADK has been associated with epileptogenesis, thus providing a rationale for adenosinemediated therapeutic intervention [10]. In previous experiments, seizure activity in kindled rats was suppressed by encapsulated adenosine releasing fibroblasts deficient for ADK [11]. However, transplanted encapsulated cells display limited long-term viability and fail to develop direct interactions with host cells. In the present study we set out to generate ADK-deficient neural cells as a tool for cellular adenosine release. To that end, we made use of embryonic stem (ES) cells. ES cells can be expanded to large numbers while maintaining their potential to differentiate into various somatic cell types. They are amenable to a wide spectrum of genetic manipulations including gene targeting. The in vitro differentiation of ES cells provides new perspectives for the generation of donor cells for transplantation therapies. Indeed, ES cells differentiate in vitro into

D.E. Fedele et al. / Neuroscience Letters 370 (2004) 160–165

clinically relevant cell types, including cardiomyocytes [14], neurons [16], and glia [4]. Following implantation into the rodent central nervous system, ES cell-derived neural precursors integrated into host tissue [5] and yielded long-term functional improvement [1]. To extend existing ES cell technology to a stem cell-based delivery system for adenosine, in the present study, both alleles of ADK were disrupted by homologous recombination in murine ES cells. As described previously [3] the gene targeting vector pAdk- was used to disrupt one allele of the Adk gene in the mEMS32 line [17] of ES cells. One correctly targeted clone was successfully used for the generation of an ADK knockout mouse [3]. The same clone was subsequently used for the genetic disruption of the second allele of Adk as follows: A total of 107 cells heterozygous for the Adk knockout were electroporated with 25 ␮g of pAdk-. Clones with a genetic disruption of both alleles of Adk were selected by using 6-methylmercaptopurine riboside (MMPR), a prodrug normally activated by ADK to form MMPR-5 -phosphate, which inhibits de novo purine nucleotide biosynthesis. Fourty hours after electroporation MMPR and guanosine were added to standard ES cell culture medium to final concentrations of 25 and 200 ␮M, respectively. After 4 days of selection a total of 4 colonies had survived. One of these colonies was propagated further and used for the experiments herein. For the analysis of homologous recombination events, DNA was extracted from the targeted ES cells and subjected to PCR with allele-specific primer sets. PCR reactions were performed under standard conditions using three primers simultaneously: o107, 5 -CTC ACT TAA GCT GTA TGG AGG TGA CCG-3 (sense primer specific for wild-type Adk), o108, 5 -AGT CAC AGA TGC ATC TGC AGA GGT GAG3 (antisense primer specific for wild-type Adk), and o109, 5 -ACT GGG TGC TCA GGT AGT GGT TGT CG-3 (antisense primer specific for targeting construct). Aqueous protein extracts from ES cells cultured in the absence of feeder cells were separated on a SDS/10%PAGE gel (30 ␮g total protein loaded), blotted onto a nitrocellulose membrane, and probed with an antiserum against mouse ADK [10]. Glial precursor cells were generated and stepwise differentiated as described [4]. They were routinely cultured on poly-ornithine-coated dishes in N3 medium, which is based on a 1:1 mixture of DMEM with Ham’s F12 supplemented with insulin (25 ␮g/ml), human apo-transferrin (100 ␮g/ml), progesterone (20 nM), putrescine (100 ␮M), sodium selenite (30 nM), penicillin (100 U/ml), and streptomycin (100 ␮g/ml). In addition, laminin (1 ␮g/ml) was added when plating the cells. To keep the cells in a proliferative state, the growth factors FGF2 (10 ng/ml) and EGF (20 ng/ml) were added daily. This medium is referred to as N3EFL medium. To promote differentiation into a stable bipotential fate, the cells were passaged and further propagated in N2FP medium consisting of basal medium as described above plus 0.15% glucose with FGF2 and PDGF added to final concen-

161

trations of 10 ng/ml each. Differentiation of cells grown in N3EFL and N2FP was induced by growth factor withdrawal. For immunostaining, glial precursor cells were plated on polyornithine-coated coverslips in appropriate medium following the procedure described above. The cells were fixed and stained with antibodies to A2B5, O4, and GFAP as described [4]. To analyze the amount of adenosine released from ES cell-derived glial cells, single cell suspensions of precursors grown in N3EFL were plated at a density of 2.5–4 × 105 cells/cm2 onto poly-ornithine coated 6-well tissue culture dishes and cultured at 37 ◦ C under 5% CO2 . For half of the dishes, 24 h after plating, the medium was changed to N2FP. For glial differentiation of cells grown in either N3EFL or N2FP, the culture dishes were withdrawn from growth factors for a period of four days under the same experimental conditions. For sample collection, the medium was replaced with 1.5 ml of fresh medium and 2 h later 200 ␮l of medium was collected and frozen at −20 ◦ C for later analysis. After collecting samples, cells on two wells from a replicate plate were counted and averaged. These data were used for normalization and quantification of adenosine release per cell number. Adenosine was quantified using an enzyme-coupled bioluminescence assay as described [11]. Statistical analysis of the results was done using a two-way ANOVA. Adenosine levels from six samples were averaged and values from Adk−/− cells were compared with the respective Adk+/+ cells at each stage of differentiation. An additional comparison was made between adenosine values from Adk−/− cells at the ES stage versus N3EFL proliferating stage, and N3EFL proliferating stage versus N3EFL withdrawn cells; this was also done for the control Adk+/+ cells. A heterozygous ES cell line with a genetic disruption of one Adk allele [3] was now used for a second round of targeting with the original Adk-specific gene targeting construct pAdk- (Fig. 1a). One MMPR resistant colony was analyzed for ADK deficiency by PCR and Western blot analysis. The wild-type specific PCR primers o107/o108 gave rise to a 640 bp band indicative of wild-type Adk which became evident in samples taken from wild-type ES cells, while DNA taken from the MMPR-resistant clone resulted in the amplification of an 840 bp knockout-specific product by primers o107/o109 (Fig. 1a and b). These PCR-based results demonstrate that the MMPR-resistant clone contains two disrupted Adk-alleles. Western blot analyses were performed on cell extracts derived from wild-type ES cells (Adk+/+ ) and the MMPRresistant Adk−/− cells, using a polyclonal rabbit antiserum raised against recombinant ADK [3]. In samples taken from wild-type cells a double band was detected in the size range of 44–46 kDa, being consistent with two alternatively spliced products (Fig. 1c). As expected, no ADK protein was detected in samples derived from MMPR-resistant Adk−/− cells (Fig. 1c). The PCR and Western Blot data demonstrate that both alleles of Adk had been disrupted, leading to ADK deficiency in these cells.

162

D.E. Fedele et al. / Neuroscience Letters 370 (2004) 160–165

Fig. 1. Bi-allelic homologous recombination of the murine Adk locus. (a) Gene maps of the murine Adk wild-type allele, targeting vector pAdk-, and targeted allele. Thick lines indicate genomic sequences of the murine Adk gene homologous to the targeting vector, while thin lines indicate external Adk sequences not present in the targeting vector. Adk-exons 6, 7, and 8 and the insertion cassette for the neomycin resistance gene (PGK-neo), are outlined as open boxes, loxP sites as vertical bars. PGK-neo is transcribed in the direction indicated by the arrow. The GFP-knock-in cassette (hatched box) for the enhanced green fluorescent protein is fused in-frame to Adk-specific exon 7, causing a disruption of the Adk gene. The black arrows indicate the location of oligonucleotide primers o107, o108, and o109 used for PCR analysis. (b) PCR analysis of the murine Adk locus. An ES cell clone lacking one Adk allele has previously been generated with the pAdk- vector [3]. To achieve a bi-allelic knockout of Adk, cells from this clone were transfected again with the same vector and selected with MMPR. Genomic DNA was amplified using the three allele specific primers o107, o108, and o109. DNA derived from wild-type ES cells (+/+) gives rise to a 640 bp band (primers o107/o108), while DNA derived from an MMPR resistant clone gives rise only to the 840 bp band (primers o107/o109) indicative for a bi-allelic genetic disruption of the Adk locus. DNA size marker: 1 kb-ladder (Life Technologies). (c) Western blot analysis of cell extracts from Adk+/+ (+/+) and Adk−/− (−/−) ES cells probed with a polyclonal rabbit antiserum raised against ADK. In samples taken from wild-type cells (+/+) a double band was detected in the size range of 44–46 kDa. No ADK specific protein was detected in samples derived from MMPR-resistant Adk−/− cells (−/−).

To exclude that ADK deficiency compromises the potential of ES cells to differentiate into glial precursor cells, we compared the glial differentiation-potential of Adk−/− ES cells with the corresponding wild-type (Adk+/+ ) ES cells in the same glial differentiation protocol. Glial precursor cells proliferating in medium containing FGF2/EGF are referred to here as “N3EFL cells” and precursors proliferating in medium containing FGF2/EGF followed by FGF2/PDGF are referred to as “N2FP cells”. Adk−/− and Adk+/+ proliferating N3EFL and N2FP cells were labeled with antibodies

against A2B5 and nestin. No significant differences were observed between the percentages of positively stained Adk−/− and Adk+/+ cells (Table 1; Fig. 2). Following growth factor withdrawal for four days, N3EFL and N2FP cells gave rise to GFAP-positive astrocytes and O4-positive oligodendrocytes (Fig. 2). Both Adk−/− and Adk+/+ cells yielded a preponderance of astrocytic cells with values ranging between 58% and 68%. In contrast, less than 6% of cells from both genotypes reached an O4-positive stage after 4 days of growth factor withdrawal (Table 1). Considering the variability be-

Table 1 Glial differentiation potential of Adk−/− vs. Adk-+/+ ES cells N3EFL

N2FP

Adk−/−

A2B5 Nestin GFAP O4

Adk+/+

Adk−/−

Adk+/+

PROL

DIFF

PROL

DIFF

PROL

DIFF

PROL

DIFF

61.0 ± 15.4 76.4 ± 4.70 nd nd

nd nd 67.4 ± 8.4 5.30 ± 2.2

64.8 ± 1.9 78.4 ± 9.0 nd nd

nd nd 65.7 ± 12.3 2.30 ± 1.50

50.7 ± 8.2 56.3 ± 0.91 nd nd

nd nd 68.3 ± 7.3 3.20 ± 1.2

48.9 ± 10.6 61.5 ± 15.7 nd nd

nd nd 58.1 ± 0.51 1.70 ± 0.61

ES cells were subjected to controlled differentiation into FGF2/EGF-induced (N3EFL) and FGF2/PDGF-induced (N2FP) glial precursors. Proliferating precursors (PROL) were induced to differentiate by a 4-day growth factor withdrawal (DIFF). Antibodies to A2B5 and the intermediate filament nestin were used to characterize the proliferating precursor cells; differentiated astrocytes and oligodendrocytes were identified using antibodies to GFAP and O4, respectively. Shown are percentages ± S.E.M. of immunolabeled cells, based on 20 high power (40×) fields per coverslip and three (1 two) independent experiments. nd: not done.

D.E. Fedele et al. / Neuroscience Letters 370 (2004) 160–165

163

Fig. 2. Glial differentiation potential of ADK-deficient ES cells. ES cell-derived glial precursors proliferating in FGF2/EGF (N3EFL PROL) and FGF2/PDGF (N2FP PROL) were induced to differentiate by a 4-day growth factor withdrawal (N3EFL DIFF; N2FP DIFF). During proliferation, both FGF2/EGF- and FGF2/PDGF-induced precursors express nestin and A2B5. Following growth factor withdrawal, they generate GFAP-positive astrocytes and O4-immunoreactive oligodendrocytes. No obvious differences in glial differentiation were noted between Adk−/− and Adk+/+ ES cells. Scale bar: 50 ␮m.

tween single experiments, no evident deficiency of Adk−/− cells for glial differentiation was noted. Thus, Adk deficiency does not affect the differentiation of ES cells into proliferating glial precursors with the potential for astrocytic and oligodendroglial differentiation. To determine whether Adk−/− ES cell-derived glial cells release adenosine in amounts considered to be of therapeutic value, we measured the adenosine levels in supernatants from undifferentiated Adk−/− and Adk+/+ ES cells and from proliferating and differentiated Adk−/− and Adk+/+ N3EFL or N2FP cells (n = 6 each). As expected, Adk−/− cells from all three differentiation stages released significantly more adenosine than the respective Adk+/+ cells (Fig. 3, p < 0.0001). It is important to note that adenosine release increased with progressive differentiation of the cells, a finding consistent with a reduced demand for adenosine incorporation into nucleic acid biosynthesis in cells with a lower proliferation rate. In

this context, adenosine release in Adk−/− cells increased significantly (p < 0.0001) from 2.6 ± 0.4 ng adenosine per 105 cells per hour (ES cells) to 11.7 ± 1.7 ng adenosine per 105 cells per hour (proliferating N3EFL cells) and 40.1 ± 6.0 ng adenosine per 105 cells per hour (N3EFL cells after growth factor withdrawal). In contrast, corresponding Adk+/+ cells released 0.17 ± 0.02 ng, 0.8 ± 0.1 ng, and 3.1 ± 0.6 ng adenosine per 105 cells per hour, respectively. Proliferating Adk−/− N2FP cells released 2.7 ± 0.8 ng adenosine per 105 cells per hour and fully differentiated Adk−/− N2FP cells after growth factor withdrawal released 11.3 ± 2.7 ng adenosine per 105 cells per hour. Corresponding Adk+/+ cells released 0.2 ± 0.0 ng and 1.9 ± 0.7 ng adenosine per 105 cells per hour, respectively. In the present contribution we describe and characterize a new stem cell-based delivery system for adenosine. Our system relies on a bi-allelic genetic disruption of the Adk

164

D.E. Fedele et al. / Neuroscience Letters 370 (2004) 160–165

Fig. 3. Adenosine release from Adk−/− ES cells and their glial progeny. Supernatant samples from the medium of ES cell-derived replicate cultures (n = 6, each) were analyzed for adenosine content. These values were normalized to the number of cells and are represented as adenosine released per 105 cells per hour. Samples were taken from wild-type (gray bars, +/+) or Adk−/− (black bars, −/−) undifferentiated ES cells, FGF2/EGF-induced glial precursors (N3EFL) and FGF2/PDGF-induced glial precursors (N2FP). Glial precursors were studied both during proliferation (PROL) and following growth factor withdrawal-induced differentiation (DIFF). Errors are given as ±S.D. Significance values have been calculated by two-way ANOVA; ∗∗∗ p < 0.0001.

gene in ES cells (Fig. 1). ES cells are useful therapeutic tools since they can be directed into multiple differentiation pathways. Most transplantation therapies with embryonic stem cells to date have focused on directed differentiation and cell replacement [1,4]. With the genetic disruption of ADK to promote adenosine release we provide the first example of engineered ES cells for drug delivery. This accomplishment further broadens the potential range of applications of ES cell transplantation in therapy. One prerequisite for a therapeutic application of ES cells is the necessity to guarantee sufficient differentiation to avoid teratoma formation by undifferentiated ES cells. It has previously been shown that wild-type ES cells can be differentiated into pure populations of glial precursor cells [4]. To assess the potential usefulness of Adk-deficient ES cells, however, it must first be demonstrated that the lack of ADK does not compromise the cells’ ability to undergo glial differentiation. Indeed, we were able to show that Adk-deficient ES cells maintained their ability to generate astrocytes and oligodendrocytes (Fig. 2, Table 1). Compared to differentiation data obtained with the ES cell lines J1 and CJ7 [4], Adk+/+ and Adk−/− mEMS32 ES cells merely yielded lower numbers of oligodendrocytes, an observation most likely due to wellrecognized differences between individual ES cell lines [12]. While further in vivo studies are required to confirm the purity and to exclude tumorigenicity of Adk−/− ES cells, their amenability to established glial differentiation protocols suggests that they may represent a suitable source of CNS transplants. In our adenosine release studies we were able to demonstrate that, after induction of differentiation into astrocytes and oligodendrocytes, cells grown in N3EFL or N2FP medium released 40.1 ± 6.0 ng and 11.3 ± 2.7 ng adeno-

sine per 105 cells per hour, respectively. Since encapsulated fibroblasts releasing between 8.4 and 19.5 ng adenosine per 105 cells per hour were previously shown to suppress seizure activity [11], adenosine levels released by Adk−/− ES cell derived astrocytes should be sufficient to provide for seizure suppression. It is important to note that adenosine is released via a ubiquitous equilibrative diffusion transporter [6], which is likely to be expressed in all cells irrespective of their state of differentiation. The direct implantation of adenosine releasing cells is expected to allow functional three-dimensional integration and long-term survival within an epileptic focus thus being advantageous compared to encapsulated cell grafts, which cannot directly interact with host cells and do not survive beyond four weeks after implantation [11]. Adk−/− ES cells may become useful tools for the treatment of various disorders, such as: (i) Epilepsy: Seizures can be suppressed by activation of adenosine A1-receptors and ventricular implants of encapsulated ADK deficient fibroblasts [11]. The integration of adenosine releasing Adk−/− ES cell derivatives might directly counteract an overexpression of ADK in epileptic hippocampus [10], and thus restore the equilibrium of adenosine. (ii) Cerebral hypoxia and ischemia: In these conditions the neuroprotective A1 receptor mediated properties of adenosine may prevent further tissue damage [8]. (iii) Chronic and neuropathic pain: Adenosine has been demonstrated to exhibit potent analgesic actions in these conditions [13]. Intrathecal implants of ES derived adenosine releasing cells thus might provide a means for pain control. (iv) Multiple sclerosis: Adenosine has been described as a mediator promoting oligodendrocyte differentiation and myelination [19]. (v) Myocardial ischemia: The cardioprotective effects of adenosine [18] could also by exploited by intracardial implants of Adk−/− ES derived cardiomyocytes. In conclusion, with the generation of Adk−/− ES cells and the demonstration of their differentiation potential and adenosine release we provide a useful donor source for transplantation therapies covering a wide range of disorders.

Acknowledgements We thank Rachel Buschwald and Michaela Segschneider for their outstanding technical support. This work was supported by grant 3100A0-100841 of the Swiss National Science Foundation and by the NCCR on Neural Plasticity and Repair, the Hertie Foundation and the Deutsche Forschungsgemeinschaft (TR-SFB 3; P.K. and O.B.). E.M.S. is a holder of a Canada Research Chair in Genetics and Behaviour.

References [1] L.M. Bjorklund, R. Sanchez-Pernaute, S. Chung, T. Andersson, I.Y. Chen, K.S. McNaught, A.L. Brownell, B.G. Jenkins, C. Wahlestedt, K.S. Kim, O. Isacson, Embryonic stem cells develop into functional dopaminergic neurons after transplantation in a Parkinson rat model, Proc. Natl. Acad. Sci. USA 99 (2002) 2344–2349.

D.E. Fedele et al. / Neuroscience Letters 370 (2004) 160–165 [2] D. Boison, Andenosine and epilepsy: from therapeutic rationale to new therapeutic strategies, Neuroscientist 10 (5) (2004), in press. [3] D. Boison, L. Scheurer, V. Zumsteg, T. R¨ulicke, P. Litynski, B. Fowler, S. Brandner, H. Mohler, Neonatal hepatic steatosis by disruption of the adenosine kinase gene, Proc. Natl. Acad. Sci. USA 99 (2002) 6985–6990. [4] O. Br¨ustle, K.N. Jones, R.D. Learish, K. Karram, K. Choudhary, O.D. Wiestler, I.D. Duncan, R.D. McKay, Embryonic stem cellderived glial precursors: a source of myelinating transplants, Science 285 (1999) 754–756. [5] O. Br¨ustle, A.C. Spiro, K. Karram, K. Choudhary, S. Okabe, R.D. McKay, In vitro-generated neural precursors participate in mammalian brain development, Proc. Natl. Acad. Sci. USA 94 (1997) 14809–14814. [6] C.E. Cass, J.D. Young, S.A. Baldwin, M.A. Cabrita, K.A. Graham, M. Griffiths, L.L. Jennings, J.R. Mackey, A.M. Ng, M.W. Ritzel, M.F. Vickers, S.Y. Yao, Nucleoside transporters of mammalian cells, Pharm. Biotechnol. 12 (1999) 313–352. [7] B.B. Fredholm, Adenosine, neuroprotection, Int. Rev. Neurobiol. 40 (1997) 259–280. [8] B.B. Fredholm, A.P. Ijzerman, K.A. Jacobson, K.N. Klotz, J. Linden, International Union of Pharmacology. XXV. Nomenclature and classification of adenosine receptors, Pharmacol. Rev. 53 (2001) 527–552. [9] N. Gouder, J.M. Fritschy, D. Boison, Seizure suppression by adenosine A1 receptor activation in a mouse model of pharmacoresistant epilepsy, Epilepsia 44 (2003) 877–885. [10] N. Gouder, L. Scheurer, J.-M. Fritschy, D. Boison, Overexpression of adenosine kinase in epileptic hippocampus contributes to epileptogenesis, J. Neurosci. 24 (2004) 692–701.

165

[11] A. Huber, V. Padrun, N. Deglon, P. Aebischer, H. Mohler, D. Boison, Grafts of adenosine-releasing cells suppress seizures in kindling epilepsy, Proc. Natl. Acad. Sci. USA 98 (2001) 7611–7616. [12] D. Humpherys, K. Eggan, H. Akutsu, K. Hochedlinger, W.M. Rideout 3rd, D. Biniszkiewicz, R. Yanagimachi, R. Jaenisch, Epigenetic instability in ES cells and cloned mice, Science 293 (2001) 95–97. [13] M.F. Jarvis, J. Mikusa, K.L. Chu, C.T. Wismer, P. Honore, E.A. Kowaluk, S. McGaraughty, Comparison of the ability of adenosine kinase inhibitors and adenosine receptor agonists to attenuate thermal hyperalgesia and reduce motor performance in rats, Pharmacol. Biochem. Behav. 73 (2002) 573–581. [14] M.G. Klug, M.H. Soonpaa, G.Y. Koh, L.J. Field, Genetically selected cardiomyocytes from differentiating embronic stem cells form stable intracardiac grafts, J. Clin. Invest. 98 (1996) 216–224. [15] K.S. Lee, P. Schubert, U. Heinemann, The anticonvulsive action of adenosine: a postsynaptic, dendritic action by a possible endogenous anticonvulsant, Brain Res. 321 (1984) 160–164. [16] S. Okabe, K. Forsberg-Nilsson, A.C. Spiro, M. Segal, R.D. McKay, Development of neuronal precursor cells and functional postmitotic neurons from embryonic stem cells in vitro, Mech. Dev. 59 (1996) 89–102. [17] E.M. Simpson, C.C. Linder, E.E. Sargent, M.T. Davisson, L.E. Mobraaten, J.J. Sharp, Genetic variation among 129 substrains and its importance for targeted mutagenesis in mice, Nat. Genet. 16 (1997) 19–27. [18] H.T. Sommerschild, K.A. Kirkeboen, Adenosine and cardioprotection during ischaemia and reperfusion—an overview, Acta Anaesthesiol. Scand. (2000) 44. [19] B. Stevens, S. Porta, L.L. Haak, V. Gallo, R.D. Fields, Adenosine: a neuron–glial transmitter promoting myelination in the CNS in response to action potentials, Neuron 36 (2002) 855–868.