Embryonic C57/Bl6 mice were used for all explant culture experiments. Mice in which the expression of GFP was driven by the rat tyrosine hydroxylase (TH) promoter 60 were used for receptor localisation studies. TH-GFP mice were on a C57/BL6 background. The genotype of adult TH-GFP mice was determined by PCR using the primers and conditions previously reported 60. Male mice heterozygous for the TH-GFP allele (TH^GFP/+ mice) were mated to wild-type (C57/BL6) females. To distinguish TH^GFP/+ from wild-type embryos, the sympathetic chain of each embryo was examined under a fluorescence microscope to determine if GFP+ cells were present; only TH^GFP/+ embryos were used. Timed pregnant mice were killed by cervical dislocation. The morning on which a copulatory plug was observed was designated E0.5. All procedures were approved by the University of Melbourne Animal Experimentation Ethics Committee.

The location of pelvic plexus primordium in E11.5 and E14.5 mice has been described previously 1115. In E11.5 mice, the plexus primordia are dorso-lateral to the distal hindgut, and in E14.5 mice, they are ventro-lateral to the distal hindgut. In P0 male mice, discrete ganglia are located on the dorsal surface of the prostate gland, and can be recognized macroscopically 1115. Connective tissue containing pelvic plexus primordia (E11.5, E14.5) or ganglia (P0) were dissected from E11.5, E14.5 and P0 mice and placed in culture medium (DMEM containing 10% foetal bovine serum, 2 mM glutamine, 0.075% sodium bicarbonate and penicillin/streptomycin sulfate solution), and were grown on collagen gels as described previously 6162. Briefly, acidic collagen solution (4 mg/ml; Upstate, CA, USA) was restored to normal osmolality with 5 × Dulbecco's modified Eagle's medium (DMEM) and normal pH with 200 mM NaOH, on ice. This solution was diluted to 1 mg/ml with culture medium. Growth factor was added to the collagen solution prior to gelling to give a final concentration of 100 ng/ml GDNF, artemin, neurturin or NGF; or 40 ng/ml NT-3 (Peprotech, NJ, USA). These concentrations are based on previous studies that have examined the responses of sympathetic and enteric neurons from embryonic and post-natal mice in vitro 616364. Each explant contained an entire ganglion. Transverse slices from E11.5 midgut and E14.5 hindgut, as well as dorsal root ganglia from E11.5, E14.5 or P0 mice, were also grown on control and growth-factor impregnated gels.

The pelvic region of TH^GFP/+ embryos was fixed overnight in 4% paraformaldehyde in 0.1 M phosphate buffer. Cryosections of 12 μm thickness were cut transverse to the neural tube and then processed for immunohistochemistry using antisera for Hu, GFP and either GFRα1, GFRα2, GFRα3, TrkA or TrkC, as shown in Table 1. Characterizations of the GFRα antibodies used have been recently described 65. Moreover, the GFRα1 antibody gives an identical pattern of staining to EGFP expression in Gfrα1^EGFP mice 66. Previous studies have shown that the TrkA antibody does not cross react with TrkB or TrkC 67. The immunogen sequence used to raise the TrkC antibody was chosen to avoid cross reactivity with TrkA and TrkB (manufacturers' information). Sections exposed to each of the secondary antisera only showed no staining.

Explants grown on collagen gels were imaged under a Leica MZ16F fluorescence stereomicroscope. Explants were immunostained using antibodies to Tuj-1 and TH. Tuj-1 immunostaining will identify all neurites and TH immunostaining will identify the cell bodies and neurites of most sympathetic neurons (see Discussion). To quantify cell and neurite outgrowth from collagen gel explants, a feathered, Otsu-algorithm thresholded mask was applied to each channel, excluding the explant, to eliminate noise from regions without outgrowth (ImageJ and Adobe Photoshop, CA, USA) as described previously 62. Three thresholded images were then produced for quantification corresponding to Tuj-1 fibres, TH-fibres and TH-cells. The total area of stained pixels was obtained, as an estimate of the total number of cells/fibres outside the explant. The quantification of Tuj-1 fibres and TH-fibres provided information about the area occupied by stained fibres outside the explant, but do not provide any information about the length of individual fibres or intensity of immunostaining. To estimate neurite outgrowth from parasympathetic neurons in the explants, the area of TH immunostaining was subtracted from the area of Tuj-1 immunostaining. The quantification of TH+ cell bodies separately from TH-fibres was possible due to TH+ cell bodies possessing an increased intensity of immunofluorescence compared to TH+ fibres, allowing their differential selection by thresholding. The quantification of TH+ cells outside of the explants provided information about the area occupied by TH+ cell bodies outside the explants, but do not provide any information about the distance of the TH+ cell bodies from the explants. TH+ cells had to be greater than one cell body diameter away from the explant to be considered outside of the explant. Measurements were made using ImageJ (NIH, USA). Results are reported as mean ± standard error of the mean. Statistical analyses were performed using one way ANOVAs followed by Dunnett's post-hoc tests. A p value of < 0.05 was deemed significant. The data for each age and each factor are based on analysis of a minimum of 3 experiments and a minimum of 7 explants.

Explants of E11.5 or E14.5 pelvic plexus primordium or P0 pelvic ganglia were grown on collagen gels, with or without 40 ng/ml NT-3, or 100 ng/ml GDNF, artemin, neurturin or NGF. After 4 days in culture, the explants were immunostained for Tuj-1 and TH; Tuj-1 immunostaining will identify all neurites and TH immunostaining will identify the cell bodies and neurites of sympathetic neurons. To estimate neurite outgrowth from parasympathetic neurons in the explants, the TH staining was subtracted from the Tuj-1 staining, which gives the area occupied by TH-negative neurites. In addition, the migration of TH (sympathetic) cell bodies was quantified. The data are summarized in Figures 1 and 2.

Transverse slices of E11.5 midgut, or E14. 5 hindgut were grown on collagen gels, with or without 40 ng/ml NT-3, or 100 ng/ml GDNF, artemin, neurturin or NGF. The data are summarized in Figure 5.

As previously shown 64, E11.5 and E14.5 gut explants grown on control gels (devoid of any neurotrophic factor) or in the presence of artemin, exhibit little or no neurite outgrowth (Fig. 5A); which is in contrast to pelvic plexus explants, which show some basal level of neurite extension in the absence of added neurotrophic factor. We observed a significant neurite outgrowth response to GDNF from E11.5 midgut slices (GDNF: 4-fold, p < 0.05; Fig. 5A). At E14.5, only neurturin induced significant neurite outgrowth (2.5-fold, p < 0.05) and there was little or no neurite outgrowth in response to GDNF (Fig. 5B). Although TrkC is expressed by a sub-population of neural crest derived cells in the gut and NT-3 has effects on differentiation 68, the effects of NGF and NT-3 on neurite outgrowth from gut slices has not been previously examined. Neither NGF nor NT-3 elicited a significant increase in total (Tuj-1+) neurite outgrowth at either E11.5 or E14.5 (Fig. 5A and 5B, respectively).

Immunohistochemical localisation of TrkA, TrkC, GFRα1, GFRα2 and GFRα3 was performed on cryosections through the pelvic region of E11.5, E14.5 and P0 TH-GFP mice. The overlap in expression of the receptors with Hu, a pan-neuronal marker, and GFP (TH) was also examined. The results are summarized in Table 2.

Of the receptor components studied, only GFRα1 was found to be expressed on Hu-negative sacral crest-derived pelvic plexus cells at E11.5. These cells are probably cells of the neural crest-derived progenitor pool, but may also be glial precursors. As all other growth factor receptors were only found on Hu+ cells, it appears that only differentiated neurons are responsive to NGF, NT-3, artemin and neurturin. GDNF-GFRα1 signaling could therefore potentially regulate the very early stages of pelvic ganglia formation, with possible actions on guidance, differentiation, proliferation or survival of the undifferentiated sacral neural crest cells.

GDNF exhibited the most potent chemotactic action on cell migration of pelvic plexus neurons of all the factors examined; GDNF stimulated migration of TH+ cells from pelvic plexus explants at E11.5, E14.5 and P0 mice. The finding that GDNF produces significant TH+ cell migration at all stages is surprising, as TH+ cells did not express detectable levels of GFRα1. There are three possible explanations: first, the migration-promoting action of GDNF is not through its canonical receptor GFRα1. GFRα2 is expressed by TH+ cells, and it has been reported that, in vitro, GDNF can actively signal through binding to GFRα2 71. However, neurturin, which most strongly activates the GFRα2 receptor, does not elicit a significant cell migration response at E11.5, and thus this possibility appears unlikely. Second, GDNF may act on the Hu-/TH- and/or Hu+/TH- sub-populations of cells, which do express GFRα1; these cells migrate in response to GDNF and then subsequently up-regulate TH expression. Finally, the number of GFRα1 receptors required to mediate a biological response might be below the detection threshold for immunohistochemical detection.

The timing of effect of the other growth factors on TH+ cell migration also gives the latter explanation further credence, as in each case the stages at which TH+ cell migration occurred correlated with the expression of the relevant receptors on the Hu+/TH- sub-population, but does not correlate with the expression of the relevant receptors on the Hu+/TH+ sub-population. Further studies are required to examine the migratory ability of Hu+/TH+, Hu+/TH- and Hu-/TH- cells.

NGF is widely expressed within the target pelvic organs of neurons of the pelvic ganglia 4972737475767778; and the neurotrophin receptors trkA and p75 are expressed by adult noradrenergic pelvic ganglion neurons 4954, as would be expected for NGF-sensitive neurons 7980. Furthermore, in mice in which both the NGF and Bax genes were knocked-out to permit the role of NGF in target innervation to be determined separately from the role of NGF in survival, there were impairments in the density of sympathetic nerves in the urinary bladder, ureter and gonads, demonstrating a role for NGF in axon pathfinding and/or axon growth and branching in these targets 81. Thus, it has been assumed that many or all pelvic noradrenergic neurons are potentially affected by endogenous NGF. However, we found NGF to be the neurotrophic factor least effective at eliciting responses from the pelvic ganglia in embryonic and newborn mice, resulting in a small but significant increase in TH+ cell and fibre outgrowth only in E14.5 mice. An earlier study using Remak's ganglia in chick, which are also derived from the sacral neural crest and are comprised of both sympathetic and parasympathetic neurons – also found a weak response to NGF throughout embryonic development with a limited period (E8–E12) of fibre outgrowth above the control level 82. This lack of effect of NGF on explants up to P0 in age contrasts with previous studies which showed NGF to stimulate sympathetic neurite outgrowth from dispersed adult rat pelvic ganglion neurons 555658. However, maturation of the pelvic ganglia continues after P0; particularly changes in the pelvic ganglia neurons due to the actions of sex hormones. Thus, it appears that the major actions of NGF occur post-natally.

Neurotrophin-3 (NT-3) is essential during sympathetic neuron development 838485, but the role of NT-3 in pelvic ganglion development has received only limited attention 4982. We found that NT-3 induces strong sympathetic (TH+) and parasympathetic (non-TH+) fibre outgrowth responses from E14.5 onwards.

The glial cell line-derived family of neurotrophic factors (GFLs) are implicated in development, maintenance, and plasticity of parasympathetic neurons 438687. Various studies have also implicated GFLs in the development or plasticity of sympathetic noradrenergic neurons 3539888990. In addition, the common GFL co-receptor, Ret, was found to be essential for survival of cholinergic sympathetic neurons in the stellate ganglion 91, and the development and/or maintenance of cholinergic sympathetic innervation of sweat glands are at least partly dependent on GFRα2 signaling 92. Thus, the GFLs are likely to play key roles in the development of the pelvic plexus, where noradrenergic sympathetic, cholinergic sympathetic and parasympathetic classes of neurons are all present.

In adult rat pelvic ganglia, GFRα1 has previously been found to be expressed by many noradrenergic and cholinergic neurons 50; however, we did not observe detectable GFRα1 immunostaining on TH+ (ie noradrenergic) neurons after E11.5. This may be because GFRα1 is only expressed by TH+ neurons after birth, or due to an interspecies difference. Despite the absence of detectable GFRα1 immunostaining by TH+ cells, we found that GDNF can stimulate outgrowth from both the TH+ and TH- classes of cells. As proposed above for the effects of GDNF-GFRα1 signaling on TH+ cell migration, some GFRα1+/TH- neurites may grow in response to GDNF, and then upregulate TH and downregulate GFRα1 during the culture period. GDNF also induced large increases in parasympathetic fibre outgrowth at E11.5 and E14.5, but had no significant effect on parasympathetic fibre outgrowth at P0. However, neurturin stimulated outgrowth of these fibres at P0, and GFRα2 was upregulated at P0 on the Hu+/TH- cells. The switch from an early importance of GDNF to a later importance for neurturin is fairly common during development of peripheral ganglia 252693.

Gfrα2-/- and Nrtn-/- mice have defects in a number of parasympathetic ganglia 879394 and the innervation of target organs by the parasympathetic component of the pelvic ganglia has also been shown to be altered. A recent study of P0-P21 Nrtn-/- mice has shown that the parasympathetic fibre deficit in each organ is quite variable and can be due to failure of initial projection to the tissue and/or maintenance of the terminal field 3. Our results also support a role for neurturin in stimulating parasympathetic (non-TH+) fibre outgrowth, as previously shown in adult rat primary neuronal culture 53, because we observed significant increases in non-TH+ fibre outgrowth at E11.5, E14.5 and P0. We also observed an effect of neurturin on sympathetic (ie TH+) fibre outgrowth, although this only occurred at E14.5. However, no major deficits have been identified in noradrenergic sympathetic neurons in Gfrα2-/- and Nrtn-/- mice 35152, so the in vivo significance of this data is unclear. We found widespread GFRα2 immunostaining on all neurons within the developing pelvic ganglia, which is in contrast to previous studies on adult mice and rat where no GFRα2 expression was found on the noradrenergic sympathetic neurons 52.

In the adult rat, most noradrenergic neurons, as well as half the penile projecting S1 DRG neurons, contained mRNA for GFRα3 50. We could not find detectable levels of GFRα3 protein in pelvic neurons using immunohistochemistry in E11.5 or P0 mice, and levels only barely above background in E14.5 mice. GFRα3 may be upregulated at post-natal stages, especially if it plays a role in the innervation of reproductive tissues, which largely occurs post-natally 3. Nonetheless, we observed artemin-induced TH+ fibre outgrowth at both E11.5 and P0, but not E14.5. This may be because the number of GFRα3 receptors required to mediate a biological response is below the detection threshold for immunohistochemical detection.

Neurotrophic factors can be involved in regulating phenotypic characteristics of neurons, including the expression of peptides and enzymes involved in neurotransmitter pathways. Studies in parasympathetic neurons have demonstrated that lack of neurotrophic support can decrease choline acetyltransferase expression and increase neuropeptide Y expression 5595. Of particular relevance to the current study are the effects of neurotrophic factors on TH synthesis: NGF has been found to enhance TH synthesis in adult rat major pelvic ganglion cultures 56; neurturin reduced the upregulation of TH expression in cultured pelvic parasympathetic neurons 53; and, GDNF has been demonstrated to downregulate expression of TH in dopamine neurons upon overexpression in vivo 96, but to upregulate the expression of TH in a neuroblastoma cell line by increasing the activity of the TH gene promoter and stabilizing TH mRNA 97. In the current study we analysed the effects of neurotrophins on the TH+ cell migration and TH+/TH- fibre outgrowth from developing pelvic plexus, but we cannot rule out the possibility that the analysis may be confounded in part due to an action of the factors on the phenotype, as well as migration and neurite outgrowth.