Identification of neurons that express ghrelin receptors in autonomic pathways originating from the spinal cord
- First Online:
- Cite this article as:
- Furness, J.B., Cho, HJ., Hunne, B. et al. Cell Tissue Res (2012) 348: 397. doi:10.1007/s00441-012-1405-9
- 319 Downloads
Functional studies have shown that subsets of autonomic preganglionic neurons respond to ghrelin and ghrelin mimetics and in situ hybridisation has revealed receptor gene expression in the cell bodies of some preganglionic neurons. Our present goal has been to determine which preganglionic neurons express ghrelin receptors by using mice expressing enhanced green fluorescent protein (EGFP) under the control of the promoter for the ghrelin receptor (also called growth hormone secretagogue receptor). The retrograde tracer Fast Blue was injected into target organs of reporter mice under anaesthesia to identify specific functional subsets of postganglionic sympathetic neurons. Cryo-sections were immunohistochemically stained by using anti-EGFP and antibodies to neuronal markers. EGFP was detected in nerve terminal varicosities in all sympathetic chain, prevertebral and pelvic ganglia and in the adrenal medulla. Non-varicose fibres associated with the ganglia were also immunoreactive. No postganglionic cell bodies contained EGFP. In sympathetic chain ganglia, most neurons were surrounded by EGFP-positive terminals. In the stellate ganglion, neurons with choline acetyltransferase immunoreactivity, some being sudomotor neurons, lacked surrounding ghrelin-receptor-expressing terminals, although these terminals were found around other neurons. In the superior cervical ganglion, the ghrelin receptor terminals innervated subgroups of neurons including neuropeptide Y (NPY)-immunoreactive neurons that projected to the anterior chamber of the eye. However, large NPY-negative neurons projecting to the acini of the submaxillary gland were not innervated by EGFP-positive varicosities. In the celiaco-superior mesenteric ganglion, almost all neurons were surrounded by positive terminals but the VIP-immunoreactive terminals of intestinofugal neurons were EGFP-negative. The pelvic ganglia contained groups of neurons without ghrelin receptor terminal innervation and other groups with positive terminals around them. Ghrelin receptors are therefore expressed by subgroups of preganglionic neurons, including those of vasoconstrictor pathways and of pathways controlling gut function, but are absent from some other neurons, including those innervating sweat glands and the secretomotor neurons that supply the submaxillary salivary glands.
KeywordsAutonomic ganglia Ghrelin Immunohistochemistry Preganglionic neurons Retrograde tracing Mouse
The ghrelin receptor was originally identified pharmacologically as the growth hormone secretagogue receptor, because its responses to artificial stimulants of growth hormone release were characterised before the endogenous ligand, ghrelin, was discovered (Howard et al. 1996; Kojima et al. 1999). Ghrelin, which is well known for increasing food intake and growth hormone release, is now also recognised as having a range of other actions (Andrews 2011). Amongst these are effects on the autonomic nervous system and on autonomically controlled organs. Some of the effects are mediated through the activation of autonomic preganglionic neurons in the spinal cord. This includes the activation of preganglionic neurons in the lumbo-sacral cord by ghrelin and by ghrelin receptor agonists (Shimizu et al. 2006; Hirayama et al. 2010), which elicit motor patterns that result in co-ordinated the emptying of the colon. Ghrelin receptor agonists also stimulate preganglionic neurons in the thoracic and lumbar regions of the spinal cord to increase blood pressure (Shimizu et al. 2006; Ferens et al. 2010a) and excitatory ghrelin receptors occur on preganglionic neurons that cause contractions of the urinary bladder (Ferens et al. 2010b). When the expression of ghrelin receptors in the spinal cord was examined by using in situ hybridisation, expression was found in the cell bodies of about 40 % of preganglionic sympathetic neurons and was observed at all levels from T3 to S2 (Ferens et al. 2010a). These data imply that ghrelin receptors are not present in all spinal autonomic pathways but that they might participate in the regulation of additional functions to those previously discovered.
In order to identify other autonomic pathways that can be influenced by ghrelin, we have used mice in which the expression of the ghrelin receptor is linked to the expression of enhanced green fluorescent protein (EGFP). In contrast to in situ hybridisation histochemistry for ghrelin receptor expression, which only reveals the cell bodies (Ferens et al. 2010a, 2010b), because little or no gene expression occurs in the nerve terminals, we have been able to reveal terminal fields by using the reporter mice. The ability to characterise subtypes of postganglionic neurons innervated by these terminals aids in the identification of the pathways of preganglionic autonomic neurons that express the ghrelin receptor.
Materials and methods
All procedures were conducted according to the National Health and Medical Research Council of Australia guidelines for animal care and were approved by the University of Melbourne Animal Experimentation Ethics Committee. Mice that expressed EGFP under control of the promoter for the ghrelin receptor were used. The EGFP reporter gene was inserted immediately upstream of the coding sequence for the receptor. Previous localisation studies with these mice had shown that EGFP was expressed in those cells in which in situ hybridisation histochemistry had identified receptor gene expression (Furness et al. 2011; Venables et al. 2011). A breeding colony was established from founder mice that were obtained from the Mutant Mouse Regional Resource Center (MMRRC) at the University of California, Davis (www.mmrrc.org/strains/30942/030942.html). Mice were genotyped by using tail snips. Genomic DNA was isolated by standard methods. The EGFP gene was identified by the polymerase chain reaction (PCR) with 5PRIME MasterMix (5PRIME, Hamburg, Germany) with forward primer 5′-GGACCTCCTCAGGGGACCAGAT-3′ and reverse primer 5′-GGTCGGGGTAGCGGCTGAA-3′ to produce a 330-bp product.
Autonomic ganglia were dissected from perfusion-fixed mice. For perfusion, reporter mice and their wild-type counterparts were anaesthetised with a mixture of xylazine (20 mg kg−1) and ketamine (100 mg kg−1), both from Troy Laboratories, Sydney, Australia, and perfused through the heart with fixative (2 % formaldehyde plus 0.2 % picric acid in 0.1 M sodium phosphate buffer, pH 7.2). Sympathetic chain ganglia (including the superior cervical and stellate ganglia), the celiaco-superior mesenteric ganglia, the inferior mesenteric ganglia and the pelvic ganglia were removed and placed in the same fixative overnight at 4°C. The next day, the tissues were trimmed and washed in phosphate-buffered saline (PBS: 0.15 M NaCl in 0.01 M sodium phosphate buffer, pH 7.2). The blocks were then placed in PBS-sucrose-azide (PBS containing 0.1 % sodium azide and 30 % sucrose as a cryoprotectant) and stored at 4°C overnight. The following day, tissue was transferred to a mixture of PBS-sucrose-azide and OCT compound (Tissue Tek, Elkhart, Ind., USA) at a ratio of 1:1 for a further 24 h before being embedded in 100 % OCT. Sections of 30 μm thickness were cut. Sections were then incubated with goat anti-EGFP raised against the full-length GFP fusion protein (Rockland Immunochemicals, Gilbertsville, Pa., USA; 0.25 μg/ml), rabbit anti-GFP (Invitrogen, Eugene, Ore., USA; 4 μg/ml) or chicken anti-GFP (Abcam, Cambridge, UK; 5 μg/ml), in order to enhance EGFP detection. In a previous study, we had shown that anti-GFP revealed the same distribution as that observed with GFP fluorescence and that no fluorescence or GFP binding was detectable in tissue from mice that did not express the transgene (Venables et al. 2011). However, notably, a component of the goat anti-GFP antiserum binds to myelin. The other anti-GFP antisera lack this artefactual binding. Sections were processed for the simultaneous localisation of EGFP and other markers: neuropeptide Y (rabbit anti-NPY; Maccarrone and Jarrott 1985), choline acetyltransferase (goat anti-ChAT; Chemicon, Boronia, Australia), vasoactive intestinal peptide (rabbit anti-VIP; Furness and Costa 1979) or synaptophysin (rabbit anti-synaptophysin; Dako, Carpinteria, Ca., USA). The sections were incubated with the primary antibodies at 4°C, overnight, before being washed and a further 1-h incubation with conjugated secondary antibodies, namely donkey anti-rabbit immunoglobulin Alexa 488 (1:1000), donkey anti-sheep immunoglobulin Alexa 594 (1:1000) and donkey anti-chicken immunoglobulin Alexa 647 (1:400), all from Molecular Probes, Eugene, Ore., USA.
Some mice were injected with the retrograde tracer, Fast Blue (2 % in 10 % dimethylsulphoxide in distilled water; Sigma-Aldrich, Sydney, Australia) under ketamine (100 mg kg−1)-xylazine (20 mg kg−1) anesthesia. Fast Blue was injected from bevelled glass micropipettes in volumes of 18 nl. Injections were made into the anterior chamber of the eye, the submaxillary gland and the forelimb muscle. For the eyes, the electrode was pushed through the sclera at a low angle until its tip was in the anterior chamber and 1-3 injections of 18 nl Fast Blue were made. Eyes were injected in four reporter mice and two wild-type mice. For submaxillary injection, the neck was opened in the midline to expose the gland. Either the right or the left gland was injected at three locations, each location receiving two 18-nl injections. Four glands were injected in reporter mice and two in wild-type mice. For the muscle injection, the skin over the upper part of the forelimb was shaved and the skin was opened so that the muscle was exposed. The limb muscles on both sides of three reporter mice were injected at three locations (2×18 nl at each site). Six or 7 days following injection, the mice were perfused with fixative and ganglia were taken and prepared as described above.
Immunoreactivity was examined by using a Zeiss Meta laser confocal microscope. Fluorophores were visualised by using a 405-nm excitation filter and 522/535-nm emission filter for Fast Blue, a 488-nm excitation filter and 522/535-nm emission filter for Alexa 488, a 568-nm excitation and 605/632-nm emission filter for Alexa 594 and a 650-nm long-pass filter for Alexa 647. Z-series images and single 1024×1024 pixel images (optical sections of 0.5 μm nominal thickness) were captured. For quantitative analysis, confocal images were taken and the cells were counted off-line. Data are presented as means±SEM. Statistical comparisons were made by using the chi-squared test and unpaired two-tailed t-tests and P<0.05 was considered significant.
When Fast Blue was injected into muscle of the forelimb, the tracer was observed in approximately equal numbers of NPY-immunoreactive and NPY-negative neurons in the stellate ganglion (Fig. 6a). The majority of both neurochemical subclasses of limb-projecting neurons had ghrelin-receptor-expressing axon terminals in close apposition (Fig. 7b).
The majority of retrogradely labelled neurons in the superior cervical ganglion following injection of Fast Blue into the anterior chamber of the eye were NPY-immunoreactive and most labelled neurons that were surrounded by ghrelin-receptor-expressing varicosities were NPY-immunoreactive (Figs. 6b, c, 7c). Of the NPY-negative neurons that projected to the eye, approximately one-third were innervated by ghrelin-receptor-expressing terminals.
Statistical analysis showed significant differences between groups of neurons that were retrogradely labelled from the salivary gland and the forelimb (chi-squared test, 2×2, P<0.05) but the differences were not significant for the neurons labelled from the eye. The numbers of neurons projecting to the salivary gland that were NPY-negative were greater than the numbers that were NPY-positive, both for neurons surrounded by ghrelin-receptor-expressing terminals and for those without terminals (t-test, P<0.001). For the forelimb, significantly more of the neurons surrounded by ghrelin-receptor-positive terminals were NPY-positive than were NPY-negative (t-test, P<0.05). Other differences were not significant (Fig. 7).
Studies in which ghrelin receptor agonists have been applied directly to the spinal cord have revealed three activated pathways: vasoconstrictor pathways (Shimizu et al. 2006; Ferens et al. 2010a), defaecation pathways (Shimizu et al. 2006) and pathways that cause contraction of the bladder detrusor muscle (Ferens et al. 2010b). Consistent with these data, in situ hybridisation histochemistry has demonstrated nerve cells expressing the ghrelin receptor gene in the autonomic preganglionic cell groups of the intermediolateral (IML) cell columns, and repetitive activation of micturition reflexes causes fos expression in preganglionic neurons of the bladder-controlling cell centres that lie in the IML and that express the ghrelin receptor gene (Ferens et al. 2010b). However, not all IML neurons express the receptor gene and another investigated pathway, namely the cardio-accelerator pathway, is not activated (Ferens et al. 2010a). The present work has been undertaken to obtain structural information concerning which spinal autonomic pathways are likely and which are unlikely to be affected by ghrelin receptor activation. Correlated light- and electron-microscope investigations in sympathetic ganglia have shown that clusters of varicose terminals that have been visualised by light microscopy to surround cell bodies make synapses with the cell bodies (Grkovic et al. 1999). Thus, if a neuron is surrounded by varicosities of a particular type, it can be concluded to be innervated by those varicosities.
The present work has confirmed that ghrelin receptors are expressed by subsets of autonomic preganglionic neurons. Terminals expressing the receptor-EGFP construct have been found in all ganglia investigated: all sympathetic chain ganglia including the superior cervical and stellate ganglia, all prevertebral ganglia and the pelvic ganglia. Terminals have also been found amongst the cells of the adrenal medulla. However, no postganglionic sympathetic neurons or adrenal or extra-adrenal chromaffin cells within the ganglia express the receptor.
Autonomic pathways lacking ghrelin receptors in preganglionic neurons
Two pathways in which preganglionic neurons lack ghrelin receptors are those to the secretory cells of the submaxillary gland and to the sweat glands. Retrograde tracing from the salivary gland has labelled large nerve cells that are not immunoreactive for NPY and smaller nerve cells, some of which are NPY-immunoreactive, in the superior cervical ganglion. Previous studies have identified the large NPY-negative neurons that are not surrounded by ghrelin-receptor-expressing axon terminals as secretomotor neurons (Lundberg et al. 1982; Gibbins 1991). Most neurons in the stellate ganglion that are immunoreactive for ChAT are cholinergic neurons that innervate sweat glands, although some cholinergic innervation of small arteries of the limbs and the periosteum has also been reported (Schäfer et al. 1997, 1998; Asmus et al. 2000). These neurons (sudomotor neurons), identified by their ChAT-IR, are not innervated by ghrelin-receptor-expressing preganglionic neurons. Clusters of neurons in the pelvic ganglia have few or no ghrelin receptor terminals in close proximity, whereas others are innervated. Previous studies have shown that parasympathetic preganglionic neurons in pathways for defaecation control and bladder contraction pass through the pelvic ganglia and are activated by ghrelin. The other major pathways that synapse in the pelvic ganglia are those to the reproductive organs (Langley and Anderson 1895; Keast 1999). Our observations suggest that preganglionic neurons in these pathways might not express ghrelin receptors.
Autonomic pathways that express ghrelin receptors in preganglionic neurons
In addition to those pathways previously functionally identified, namely to blood vessels, the colorectum and bladder, several other pathways have been identified structurally by the present investigation, which also provides additional information about pathways to the vasculature.
Subgroups of neurons in the celiaco-superior mesenteric ganglion complex provide major inputs to mesenteric and gastrointestinal blood vessels, motility controlling neurons in the enteric nervous system of the stomach, small intestine and part of the large intestine, and enteric neurons that control water and electrolyte transport in the small intestine and upper large intestine (Lomax et al. 2010). The sympathetic neurons that control motility and fluid transport are clumped in the ganglia; they receive synaptic inputs from VIP-immunoreactive intestinofugal neurons and can be identified by their VIP innervation (Macrae et al. 1986; Lindh et al. 1988). We have found that neurons surrounded by VIP terminals are also innervated by ghrelin receptor terminals, indicating that both the motility-controlling and fluid-transport-controlling pathways involve preganglionic neurons that express the receptor. Ghrelin affects the movements of the proximal gastrointestinal tract (Tack et al. 2006; Fujimiya et al. 2010) but whether the activation of sympathetic innervation plays a part in these effects is not as yet resolved. Other less numerous cell populations in these ganglia, such as those innervating the spleen and pancreas, will need to be specifically identified, for example, by retrograde transport, to determine whether their preganglionic inputs express ghrelin receptors. The fasting-induced release of ghrelin from the stomach is dependent on sympathetic pathways and is prevented by noradrenaline depletion following reserpine treatment (Zhao et al. 2010). The secretion is mediated through β1 adrenergic receptors. Currently, the preganglionic sympathetic neurons of this pathway have not been specifically identified.
Another pathway in which preganglionic neurons express ghrelin receptors is that to the anterior of the eye. The major target of this pathway is the dilator muscle of the iris, although a few fibres also innervate blood vessels of the iris (Vidovic et al. 1987). Ghrelin relaxes the dilator muscle by what appears to be a direct action (Rocha-Sousa et al. 2006).
The current study adds to knowledge of the vasoconstrictor innervation of blood vessels by pathways in which the preganglionic neurons express ghrelin receptors. Ghrelin or non-peptide ghrelin receptor agonists applied to the spinal cord have been previously found to raise blood pressure (Shimizu et al. 2006; Ferens et al. 2010a), but the vascular beds that are constricted to cause this increase are not known, although the greatest effects have been obtained by agonist application at T9 and T10, with pressor effects being detected with drug application from T8 to L3. In the current study, we have injected Fast Blue into the muscle of the forelimb in which sympathetic nerve terminals innervate the arteries supplying the muscle. Axons in nerves passing through the muscle might also be labelled. The vasoconstrictor neurons express NPY (Lundberg et al. 1982; Gibbins 1991). We have found that 90 % of the NPY-immunoreactive neurons that are retrogradely labelled are innervated by ghrelin-receptor-expressing preganglionic neurons. In the celiac ganglion of the guinea-pig, about half the neurons are vasoconstrictor neurons (Macrae et al. 1986). This has not been directly investigated in mouse, a species in which a chemical distinction between vasomotor and other neurons of prevertebral ganglia has not been established and in which more than the vasoconstrictor neurons are NPY-immunoreactive. However, because almost all neurons are innervated by ghrelin receptor terminals in these ganglia, we can assume that this includes the cell bodies of vasoconstrictor neurons. Thus, the pressor effect of activating ghrelin receptors in the spinal cord probably involves the constriction of arteries that supply the abdominal viscera and skeletal muscle. We have also found an innervation of the adrenal medulla by ghrelin-receptor-expressing neurons. Activation of these neurons would contribute to blood pressure increases.
Ghrelin receptors are expressed by subsets of preganglionic neurons of autonomic pathways originating from all levels of the spinal cord but are not present in postganglionic neurons. Expression is restricted to specific pathways, including the vasoconstrictor and pupillodilator pathways and those pathways affecting gastrointestinal and bladder function. On the other hand, expression is absent from pathways to the sweat glands and the secretomotor neurons that innervate the submaxillary salivary glands.
We thank Dr. Colin Anderson and Mr. David Gonsalves for their advice and assistance with the Fast Blue injections and for their comments on the manuscript and Dr. Trung Nguyen for advice on the statistics.