Journal of Neurochemistry, 2006, 99, 343–352
BMPs, FGF8 and Wnts regulate the differentiation of locus coeruleus noradrenergic neuronal precursors Pontus C. Holm,*,1 Francisco J. Rodrı´guez,*,2 Julianna Kele,* Gonc¸alo Castelo-Branco,*,1 Jan Kitajewski and Ernest Arenas* *Laboratory of Molecular Neurobiology, MBB, Karolinska Institutet, Stockholm, Sweden Department of Pathology, Columbia University, New York, New York, USA
Abstract In the present study, we investigated the involvement of rhombomere 1 patterning proteins in the regulation of the major noradrenergic centre of the brain, the locus coeruleus. Primary cultures of rat embryonic day 13.5 locus coeruleus were treated with fibroblast growth factor-8, noggin and members of the bone morphogenetic and Wnt protein families. We show that bone morphogenetic proteins 2, 5 and 7 increase and noggin decreases the number of tyrosine hydroxylase-positive locus coeruleus neurons. Interestingly, from all Wnts expressed in the first rhombomere by embryonic day 12.5 in the mice, we only found expression of wnt5a mRNA in the vicinity of the locus coeruleus. In agreement with this finding, from all Wnts studied in vitro, only
Wnt5a increased the number of tyrosine hydroxylase-positive neurons in locus coeruleus cultures. Finally, we also found that fibroblast growth factor-8 increased the number of tyrosine hydroxylase-positive cells in locus coeruleus cultures. Neither of the identified factors affected the survival of tyrosine hydroxylase-positive locus coeruleus noradrenergic neurons or the proliferation of their progenitors or neurogenesis. Instead, our results suggest that these patterning signals of rhombomere 1 may work to promote the differentiation of noradrenergic progenitors at later stages of development. Keywords: bone morphogenetic protein, fibroblast growth factor, locus coeruleus, noradrenergic neurons, Wnt. J. Neurochem. (2006) 99, 343–352.
The locus coeruleus (LC) is localized in the lateral walls of the fourth ventricle, in the anterior segment of the hindbrain, in rhombomere 1. This nucleus is the main noradrenergic centre of the brain and the main source of noradrenergic innervation in the CNS. The early development of noradrenergic neurons of the LC requires the establishment of an anterior–posterior pattern by the midbrain–hindbrain organizer. Some of the known genes involved in the process include, ﬁbroblast growth factor-8 (FGF-8; Ye et al. 1998; Guo et al. 1999b), Wnt1 (McMahon and Bradley 1990; Thomas and Capecchi 1990), engrailed-1 (Wurst et al. 1994; Danielian and McMahon 1996), engrailed-2 (Hanks et al. 1995) Pax2 (Favor et al. 1996) and Pax5 (Urbanek et al. 1997). Other types of genes involved include factors responsible for the establishment of dorso– ventral pattern. In that respect, the development of ventral phenotypes require ‘sonic hedgehog’ (Hynes and Rosenthal 1999), a factor derived from the ﬂoor plate. Instead, the development of dorsal phenotypes, including that of noradrenergic neurons, require bone morphogenetic proteins
(BMPs; Lo et al. 1998; Guo et al. 1999a,b; Lo et al. 1999; Vogel-Hopker and Rohrer 2002), which are expressed in the dorsal rhombomere 1 and the adjacent choroid plexus (Furuta et al. 1997; Solloway and Robertson 1999; VogelHopker and Rohrer 2002). In addition to these two types of signals, the induction of speciﬁc cell types require the Received March 6, 2006; revised manuscript received May 23, 2006; accepted June 8, 2006. Address correspondence and reprint requests to Ernest Arenas, Laboratory of Molecular Neurobiology, MBB, Karolinska Institutet, Scheeles va¨g 1, A1:2, 171 77 Stockholm, Sweden. E-mail: [email protected]
1 The present address of Pontus C. Holm and Gonc¸alo Castelo-Branco is the Department of Neuroscience, Karolinska Institutet, Retzius va¨g 8, A2:2, 171 77 Stockholm, Sweden. 2 The present address of Francisco J. Rodriguez is Laboratorio de Neurologı´a Molecular, Hospital Nacional de Paraple´jicos, 45071 Toledo, Spain. Abbreviations used: BMP, bone morphogenetic protein; E13.5, embryonic day 13.5; FGF, ﬁbroblast growth factor; GDNF, glial cell-linederived neurotrophic factor; LC, locus coeruleus; NA, noradrenergic; TH, tyrosine hydroxylase.
2006 The Authors Journal Compilation 2006 International Society for Neurochemistry, J. Neurochem. (2006) 99, 343–352
344 P. C. Holm et al.
expression of transcription factors that turn on speciﬁc genetic programmes and provide the cells with a deﬁned neurotransmitter identity. For instance, for their development, noradrenergic neurons from the locus coeruleus require the basic helix–loop–helix transcription factor, mash1 (Hirsch et al. 1998; Lo et al. 1998), and the paired homeodomain proteins, phox2a (Lo et al. 1999) and phox2b (Tsarovina et al. 2004), as demonstrated by the lack of locus coeruleus noradrenergic neurons in null mutant mice for these genes. In addition, the homeobox transcription factor tlx3 (Qian et al. 2001; Hornbruch et al. 2005), and the orphan receptor ear2 (Warnecke et al. 2005), have been described to be necessary for adequate LC development. With regard to the trophic requirements of LC noradrenergic (NA) neurons, in vivo experiments on adult lesion models have shown that both NT-3 (Arenas and Persson 1994) and glial cell-line-derived neurotrophic factor (GDNF; Arenas et al. 1995) are survival factors for adult LC NA neurons. However, mice with targeted deletion of the nt-3 gene, the gdnf gene or both the nt-3 and the gdnf genes have shown no loss of NA neurons at birth (Holm et al. 2003). Instead, we reported a crucial role for the tyrosine kinase receptor TrkB and its ligands BDNF and NT-4 for the differentiation and survival of developing noradrenergic LC neurons both in vitro and in vivo (Holm et al. 2003). In the present study, we set out to determine whether epigenetic signals responsible for the patterning of rhombomere 1 could be involved in regulating the survival and differentiation of LC NA neurons. Our results show that several BMPs (2, 5 and 7), Wnt5a and FGF8 are able to increase the number of tyrosine hydroxylase (TH)positive LC neurons in embryonic day 13.5 (E13.5) primary cultures, suggesting that signals that earlier pattern rhombomere 1 are, at latter stages, involved in controlling the differentiation of LC noradrenergic neurons.
Experimental procedures Locus coeruleus cell culture and treatments LC primary cultures were performed as previously described (Holm et al. 2003). In brief, pregnant Sprague–Dawley rats (B&K Universal AB, Sollentuna, Sweden) were killed at E13.5 and embryos were removed for microscopic dissection of the whole proximal rombencephalic ring, between the distal part of the mesencephalic ﬂexure and the proximal part of the pontine ﬂexure (Specht et al. 1981a,b), including the LC. The tissue was gently dissociated with a ﬁre-polished Pasteur pipette and plated on poly Dlysine coated 12-well plates (BD Falcon, Franklin Lakes, NJ, USA) in N2 media at a density of 1.25 · 105 cells per cm2 (5 · 105 cells/ well). Factors were added to the wells just before the cells were plated. BMP2, -4, -5, -6 and -7, GDF6 and FGF8b were all obtained from R & D Systems (Minneapolis, MN, USA). Solutions used in tissue culture were sterilized by ﬁltration through 0.22-lm ﬁlters (Millipore Corporation, Bedford, MA, USA). Cultures were grown for 6 days (except for where stated otherwise) at 37C in a water-
saturated 5% CO2/95% air atmosphere incubator without changing media or supplementing the factors. Conditioned media preparation, characterization and purification Rat ﬁbroblast lines stably over-expressing haemagglutinin-tagged Wnt1a, -3a, -5a, -7a (Shimizu et al. 1997) and Fz8-CRD (Hsieh et al. 1999; Dann et al. 2001; B1A) and Wnt2 (Baﬁco et al. 1998; 3T3) were grown in standard complete media (Dulbecco’s modiﬁed Eagle’s medium + 10% fetal bovine serum) supplemented with 100 lg/mL G-418. Collection and partial puriﬁcation of conditioned media was performed as previously described (Castelo-Branco et al. 2003). In brief, cells were re-plated at low density in complete media and allowed to reach 50–75% conﬂuency, at which point cells were washed and media was replaced with serum-free N2 for 24 h. Individual lots of conditioned media were loaded onto CentriconPlus 80 columns (Millipore) and concentrated via centrifugation according to the manufacturer’s instructions. Following concentration, aliquots were frozen at )80C after a sample was taken for determination of protein content and western blot analysis. Partially puriﬁed B1A, 3T3, Wnt and Fz8-CRD conditioned media had a ﬁnal total protein concentration of 25 lg/mL. Volumes used in culture conditions were 5 lL/mL of Fz8-CRD and 3, 10 or 30 lL/mL of the others. Immunocytochemistry and cell counts of primary cultures Cultures were ﬁxed in 4% paraformaldehyde in phosphate-buffer saline with pH 7.4 and then incubated with primary antibody, diluted in phosphate-buffered saline containing 1% bovine serum albumin and 0.3% Triton X-100, at 4C overnight. After washing, cultures were incubated for 1–2 h with biotinylated secondary antibody, diluted 1 : 500 in the same dilution buffer. Immunostaining was visualized with the ABC immunoperoxidase kit (Vector Laboratories, Burlingame, CA, USA), using grey (SG) substrate. Used as primary antibodies were mouse anti-TH (DiaSorin, Salaggia, Italy), diluted 1 : 1000, mouse anti-ki67 (Abcam, Cambridge, UK), diluted 1 : 800, mouse anti-TuJ (Sigma, St Louis, MO, USA), diluted 1 : 250, and mouse anti-BrdU (Dako, Carpinteria, CA, USA), diluted 1 : 50. Secondary antibodies were horse anti-mouse IgG (Vector Laboratories) following the TH, TuJ1, and BrdU antibodies and horse antimouse IgM (Vector Laboratories) following the ki67 antibody. All the TH- and Ki67-positive cells in each immunostained culture well were counted. In BrdU and TuJ1 immunostained wells, 20 microscope ﬁel...