|Home | About | Journals | Submit | Contact Us | Français|
Vertebrate α–bungarotoxin-like molecules of the Ly-6 super family have been implicated as balancers of activity and survival in the adult nervous system. To determine whether a member of this family could be involved in the development of the avian ciliary ganglion, we identified 6 Gallus genes by their homology in structure to mouse lynx1 and lynx2. One of these genes, an ortholog of prostate stem cell antigen (PSCA), is barely detectable at embryonic day 8, prior to neuronal cell loss in the ciliary ganglion, but increases over 100-fold as the number of neurons begins to decline between E9 and E14. PSCA is highly expressed in chicken and mouse telencephalon and peripheral ganglia and correlates with expression of α7-containing nicotinic acetylcholine receptors (α7-nAChRs). Misexpressing PSCA prior to cell death in the ciliary ganglion blocks α7-nAChR activation by nicotine and rescues the choroid subpopulation from dying. Thus, PSCA, a molecule previously identified as a marker of prostate cancer, is a member of the Ly-6 neurotoxin-like family in the nervous system, and is likely to play a role as a modulator of α7 signaling induced cell death during development.
Nicotinic signaling has been implicated in controlling programmed cell death during the development of the nervous system. By far the most effective means of rescuing spinal cord motor neurons from dying is to treat embryos with nicotinic antagonists that block neuromuscular transmission such as d-tubocurarine (dTC) or alpha-bungarotoxin (αbtx) (Pittman and Oppenheim, 1979). In contrast, in the avian ciliary ganglion, up to 90% of the neurons in the avian ciliary ganglion are rescued by antagonists of α7-containing nicotinic acetylcholine receptors (nAChRs) such as αbtx and alpha-methyllycaconitine (MLA; Bunker and Nishi, 2002; Hruska et al, 2007); however, non-selective nAChR antagonists that completely block ganglionic transmission exacerbate cell death (Wright, 1981; Meriney et al., 1987; Maderdrut et al., 1988). In cell culture, nicotine promotes survival of ciliary ganglion neurons, but it does so in the absence of trophic support (Pugh and Margiotta, 2000). Thus, the contribution of nicotinic signaling to neuronal survival during development is complex.
Endogenous prototoxins that are homologous to snake venom neurotoxins modulate signaling through nAChRs. These molecules are members of the Ly6/neurotoxin (lynx) family characterized by their cysteine-rich motif that predicts their folding into the typical three-fingered loop structure of α–bungarotoxin (Miwa et al., 1999; Chimienti et al., 2003). Members of this family that are GPI-linked and expressed in the nervous system are lynx1, lynx2, Ly6H, (Horie et al., 1998; Miwa et al., 1999; Dessaud et al., 2006); others are secreted and expressed in non-neuronal cells, e.g., SLURP-1 and SLURP-2; (Chimienti et al., 2003; Tsuji et al., 2003). Mouse lynx1 binds to and alters the kinetics of nAChRs (Ibanez-Tallon et al., 2002). Furthermore, cortical neurons of mice lacking lynx1 have an enhanced sensitivity to nicotine that potentiates excitotoxicity, while aging lynx1 null mice exhibit exacerbated neurodegeneration in the brain that is enhanced by nicotine and attenuated by loss of nAChRs (Miwa et al., 2006).
In the present study we determined whether the expression of endogenous lynx-like molecules could play a role during the development of the avian ciliary ganglion. The ganglion contains two types of neurons, ciliary and choroid, and both receive cholinergic innervation that is detectable at embryonic day (E) 5.5 and complete by E8 (Landmesser and Pilar, 1972). Over half of the ciliary and choroid neurons are lost between E8 and E14 (Landmesser and Pilar, 1974). By E8, all neurons express homomeric α7-nAChRs and heteromeric α3* nAChRs containing α3, α5, β4 subunits (Blumenthal et al., 1999; McNerney et al., 2000) and sometimes β2 (Conroy and Berg, 1995). Using mouse lynx1 and lynx2 we screened for homologous molecules in the chicken and identified 6 genes. Of these, one is induced in the ciliary ganglion during the period of cell loss and is the chicken ortholog of human prostate stem cell antigen (PSCA). PSCA is developmentally regulated and its expression reduces α7-nAChRmediated increases in intracellular Ca2+ and promotes survival of choroid neurons. We conclude that PSCA, a molecule previously identified as upregulated in prostate cancer, is also a lynx-like prototoxin molecule that functions in the developing nervous system.
Mouse lynx1 and lynx2 protein sequences were blasted against chicken TIGR gene indices (http://www.tigr.org/tdb/tgi/) and we chose 6 sequences that fit the following criteria: 1) having the Ly6 domain as the only functional domain; 2) matching the exon-intron structures of Lynx1 and other prototoxins that have the Ly6 domain. The accession numbers of the sequences that we discovered are given on Supplemental Table 1. Based on predicted sequences, sequence specific primers were used to identify the expression profile of above lynx related transcripts in the ciliary ganglion at E8 and E15 (for primer sequences, see Supplemental Table 1).
Tissues were isolated at the indicated stages of development and rapidly frozen on dry ice. Total RNA was extracted by using TRI Reagent (Molecular Research Center, Inc., Cincinnati, OH). The cDNA was reverse-transcribed from 2 µg of total RNA by using oligo-dT and Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA). The full length PSCA was obtained by PCR using forward primer (5’ CCATGGTAATGAAGGTTTTCTTCATCCTCC 3’) that attached an Nco1 restriction site to the 5’ end of a PSCA sequence and reverse primer (5’ GGATCCCTTCACAGTCTGTTGTTCAGG 3’) that added a BamH1 restriction site to the 3’ end of the PSCA sequence. The 369 base pair PCR product was cloned into Nco1 and BamH1 cloning sites on pSlax13Nco vector and sequencing confirmed that it was the accurate, full-length chicken PSCA sequence.
Tissues were isolated at the indicated ages and rapidly frozen on dry ice. Total RNA was extracted and cDNA was synthesized as described above. The expression profiles of α7 and PSCA were assessed on the 7500 Fast Real-Time PCR System (Applied Biosystems, Foster City, CA) using TaqMan probes and primers designed using a website at Whitehead Institute at MIT http://frodo.wi.mit.edu/primer3/ (see Supplemental Table 2). Probes were labeled with 6 FAM reporter dye at their 5’ ends and black hole quencher (BHQ) at their 3’ ends. The constitutive gene control used to normalize gene expression was chicken S17 ribosomal binding protein (Darland et al., 1995). Relative transcript expression and number was determined using Sequence Detection Software (SDS) version 1.4. PSCA expression in mouse tissues was quantified and normalized to β-actin using Assays on Demand from Applied Biosystems (PSCA: Mm00452908_m1; β-actin: Mm00607939_s1).
The PSCA sequence was cloned into the Slax13NCO1 shuttle vector using 5’ Nco1 and 3’ BamH1 sites (Morgan and Fekete, 1996); the insert was removed from Slax13NCO1 by cutting with Cla1, and cloned into the avian retroviral vector RCASBP(A) (Federspiel and Hughes, 1997). Infective RCASBP(A)-PSCA viral particles were generated by transfecting DF-1 chicken fibroblast cells with 800 ng of RCASBP(A)-PSCA plasmid using Mirus TransIT-LT transfection reagent (Mirus Bio Corporation, Madison, WI). Conditioned media containing viral stocks collected from DF-1 cells were concentrated approximately 20-fold by ultracentrifugation at 90,000×g at 4°C for 3 hr (Morgan and Fekete, 1996). Concentrated stocks were titered by limiting dilution and infectivity of cells as measured by staining with p27gag antibody. Stocks with concentration of >108 infectious particles/ml were used for in vivo injection. Viral particles (60–120 nL) were injected into the mesencephalic enlargement of the neural tube of Hamburger/Hamilton stage (St.) 8–9 or St. 10–13 embryos using a Drummond Nanoject microinjector (Drummond Scientific, Broomall, PA). The shells were sealed with a glass coverslip and sterile vacuum grease and incubated at 37°C to the desired stage.
Acutely isolated ciliary ganglion neurons were loaded with Fura-2 AM (Invitrogen) dissolved in DMSO at final concentration of 5 µM with 2% Pluronic F-127 (Invitrogen). Neurons were loaded at room temperature for 30 min in the dark. Calcium signals were recorded by exposure to alternating wavelength (340 and 380 nm, 50 ms) generated by a Xenon light source and Lambda DG-4 ultra high-speed wavelengths switcher (Sutter Instruments). Fluorescent responses were recorded using an Orca-ER digital camera (Hammatsu). Paired 340/380 ratio images were acquired at 4s intervals with Metaflour 5.0r5 software (Molecular Devices, Sunnyvale, CA). Drugs were dissolved in chicken physiological buffer (145 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl2, 5.4 mM CaCl2, 5 mM glucose, 13 mM HEPES, pH 7.4). Voltage gated sodium and calcium channels were blocked when nicotine was applied by supplementing the perfusion medium with 600 nM tetrodotoxin (Tocris, Ellisville, MO) and 200 µM cobalt chloride (Sigma-Aldrich, St. Louis, MO), respectively. 10 µM nicotine (Sigma-Aldrich) was applied for 20 sec to activate nAChRs. α7-nAChRs were inhibited by perfusing the neurons with 50 nM α-methyllycaconitine citrate hydrate (MLA; Sigma-Aldrich) for 1min or pre-incubating with 50 nM αbtx (Sigma-Aldrich) for 30 min at 25°C. Upon completion of these experiments, the extent of dye loading was determined by activating voltage-gated calcium channels with high potassium perfusion solution (25 mM KCl) in the absence of TTX or cobalt chloride. After the initial recordings were performed, background was subtracted from every image acquired and new ratios were calculated using the Metafluor 5.0r5 software. The ratios were then exported into the Microsoft Excel spreadsheet for calculations.
Ciliary ganglia and brainstems from E14 embryos were harvested, fixed in Zamboni’s for 48 hours at 4°C, washed, then equilibrated in 30% sucrose at 4°C. Tissue was embedded in Microm cryo-embedding compound (Richard Allen Scientific, Kalamazoo, MI), sectioned on a Microm HM 560 cryostat (Richard Allen Scientific) at 30 µm (for immunohistochemistry) and at 20 µm for in situ hybridization and collected on Superfrost Plus slides (Fisher Scientific, Pittsburgh, PA). Sections were post-fixed in Zamboni’s vapors for 15 min at 37°C, submerged in Zamboni’s fixative for additional 15 min at 25°C, washed in PBS and blocked. Primary antibodies were incubated overnight at 4°C and secondary antibodies were incubated 2 hr at room temperature.
In situ hybridization was performed as described by the D. Anderson laboratory at Caltech (http://wmc.rodentia.com/docs/Big_In_Situ.html) with the following modifications: sections collected on slides were treated with 100 µg/ml of proteinase K (Sigma-Aldrich) for 10 min at room temperature; hybridization was performed with digoxigenin-labeled UTP (Roche) -labeled sense or antisense riboprobes synthesized using an RNA transcription kit (Ambion) from a full length Gallus clone gcag0014c.j.08 corresponding to the cDNA we cloned, which was obtained from the French National Institute for Agricultural Research Animal Genomics Collection (http://www.international.inra.fr); and riboprobes were subjected to alkaline hydrolysis until average transcript sizes were 300–400 bp.
Mouse monoclonal 39.4D5 (Developmental Studies Hybridoma Bank (DSHB)), which recognizes Islet-1 and Islet-2, transcription factors expressed in all ciliary ganglion neurons throughout development (Lee et al., 2001) at 1:100 dilution of the culture supernatant prepared in the Nishi Lab from hybridomas obtained from DSHB; mouse anti-Hu C/D (Invitrogen), which recognizes a neuron-specific RNA binding protein (Marusich and Weston, 1992; Lee et al., 2001) at 1:250 dilution of the culture supernatant; rabbit anti-p27gag (Charles River SPAFAS, North Franklin, CT), which recognizes avian sarcoma gag p27 (Wang et al., 1976) at 1:1000; rat anti-somatostatin (Product #: YMC1020, Accurate Chemical & Scientific Corp., Westbury, NY) diluted 1:100. Sheep anti-digoxigenin (Roche Applied Science, Indianapolis, IN) was used at 1:500 and visualized with goat anti-sheep coupled to alkaline phosphatase (1:1000; Roche).
biotinylated anti-mouse (Vector Laboratories) at 1:250; biotinylated anti-rabbit (Vector Laboratories, Burlingame, CA) at 1:250; goat anti-mouse Cy3 (Jackson Immuno Research) at 1:750; goat anti-rabbit Alexa 488 (Invitrogen) at 1:750; and goat anti-rat Cy3 (Jackson Immuno Research, West Grove, PA) at 1:750. Images of ciliary ganglia and midbrain were acquired with 2× objective using Nikon C1 confocal scanner (Nikon Instruments, Melville, NY) attached to a Nikon Eclipse E800 microscope (Micro Video Instruments, Avon, MA).
Serially sectioned ciliary ganglia (cut at 30 µm) were prepared for designed-based stereology as previously described (Lee et al., 2001; Bunker and Nishi, 2002). Islet-1/2 positive nuclei (representing all neurons) together with somatostatin-positive cell bodies (representing all choroid neurons) were counted using the Optical Fractionator Probe of Stereo Investigator (MBF Biosciences, Williston, VT) in conjunction with a Nikon Optiphot 2 microscope with a Hitachi HVC20 camera, Heidenhahn focus encoder, and a motorized, computer-driven X, Y, Z stage (all microscope attachments provided by MBF Biosciences). To avoid inaccuracies caused by cutting artifacts and double counting between adjacent sections, an upper guard of 4 µm and lower guard of 7 µm were used (Bunker and Nishi, 2002). Spacing between sampling sites (grid size) was set such that 13–15 sampling sites were counted per section, which yielded 100–300 objects per each ciliary ganglion. The number of ciliary neurons was calculated by subtracting the number of somatostatin-positive neurons from the total number of neurons per ciliary ganglion (Bunker and Nishi, 2002).
Since mouse lynx1 interacts with α7-nAChRs from chicken (Ibanez-Tallon et al., 2002), we used the amino acid sequences of mouse lynx1 and lynx2 to search a chicken expressed sequence tag (EST) database for proteins that had the Ly6 domain as the only functional domain and whose genes had similar exon-intron structures to those of lynx1 and other toxins that have the Ly6 domain. Six sequences that matched these criteria were identified (Fig. 1A). All 6 sequences are cysteine-rich and 10 of the cysteine residues are aligned across all molecules. The sequences also encode an N-terminal signal sequence and a C-terminal consensus sequence for addition of a glycosyl-phosphotidylinositide linkage to the membrane. When analyzed by SMART (Simple Modular Architecture Research Tool; http://smart.embl-heidelberg.de/), all molecules are also predicted to fold into a three-fingered coiled structure similar to that of mouse lynx1 and the cobra venom neurotoxins (Gumley et al., 1995). This striking structural homology suggests that these chicken molecules can interact with nicotinic acetylcholine receptors as shown for mouse lynx1 (Ibanez-Tallon et al., 2002).
To determine if any of the 6 EST transcripts were expressed in a developmentally regulated pattern in the ciliary ganglion, we used sequence-specific primers (Supplemental Table 1) to amplify cDNA from ciliary ganglia collected at E8, while cell death is occurring, and E14, at which time cell death has ceased (Lee et al., 2001). Three sequences, ch3Ly, ch5Ly and ch6Ly were expressed in E14 ganglia; however, ch3Ly and ch5Ly were also expressed at E8 (Fig. 1B). In contrast, no expression of ch6Ly is detectable at E8 (Fig. 1B). Identity of the amplified gene product as ch6Ly was confirmed by sequencing. The remaining transcripts are not expressed in the embryonic ciliary ganglion (data not shown). Ch6Ly encodes a pro-protein of 122 amino acids, which corresponds to a molecular weight of 11,160 Da. The mature GPI-linked protein is 8,051 Da. When ch6Ly is used to search Entrez, it matches mouse prostate stem cell antigen (PSCA), with which it shares 40% amino acid identity and 80% homology (Fig. 2A). Like ch6Ly, the open reading frame of mouse PSCA is similar to other Ly6 superfamily members in the alignment of cysteine groups and contains the Ly6 domain. In addition, the genomic structure of mouse PSCA is identical to that of ch6Ly, mouse lynx1, mouse ly6h, and αbungarotoxin, with conserved intron/exon breaks (Fig. 2B). Thus, ch6Ly is likely to be the chicken ortholog of prostate stem cell antigen (PSCA), which is a Ly6 family member whose expression becomes upregulated in prostate tumors (Reiter et al., 1998).
To determine the specificity of chicken PSCA expression, we quantified transcripts in a variety of tissues in E14 chicken embryos using real-time PCR and normalized these levels to chicken ribosomal binding protein s17 (CHRPS), a constitutively expressed housekeeping gene that is abundant in all tissues. Expression of PSCA in selected tissues was determined relative to that of heart, which contained barely detectable levels (Fig. 3A). Levels of PSCA transcripts in pectoral muscle, liver, ovary and testes (data not shown) are comparable to that of heart. In the nervous system, the cerebellum also has very low levels; however, the telencephalon has 5-fold greater levels of PSCA mRNA than cerebellum. The highest levels of PSCA occur in autonomic ganglia: the ciliary ganglion, which is parasympathetic, contains more than 20-times that of heart, and paravertebral lumbar sympathetic ganglia contain 10-times more PSCA. Dorsal root ganglia also express higher levels of PSCA mRNA than heart, but considerably less than sympathetic or ciliary ganglia.
Since our observation that PSCA is enriched in telencephalon was contrary to the original report that PSCA could not be detected in northern blots of brain mRNA (Reiter et al., 1998), we determined whether adult mouse tissues exhibit an expression pattern of PSCA mRNA similar to that of chicken embryos. We isolated neural and non-neural tissues and quantified relative abundance of mouse PSCA normalized to β-actin using real time PCR, which is more sensitive than the northern blot approach previously used. As in late stage chicken embryos, the nervous tissues isolated from adult mice contain significantly higher levels of PSCA than non-neural tissues, with the highest levels found in the superior cervical ganglion, which is sympathetic (Fig. 3B). In fact, all the tissues that exhibit high levels of PSCA in the chicken embryo also express high levels of α7-nAChRs (Fig. 3C). Thus, PSCA is expressed at very high levels in neural tissues both in mammals and avian species, and its level of expression correlates with the expression of the α7-nAChR subunit (Fig. 3C, 3D), suggesting a relationship between PSCA and α7 signaling.
In ciliary ganglia, which express some of the highest levels of α7 nAChRs per neuron, α7-nAChRs have previously been implicated in inducing cell death (Bunker and Nishi, 2002; Hruska et al, 2007). If PSCA attenuates cell death by modulating activation of α7 subunit containing nAChRs, then its expression should be low when cell death commences, and it should be upregulated as neurons in the ganglion extend axons to the periphery and initiate synaptogenesis. Accordingly, we used real time PCR to quantify PSCA transcripts in ciliary ganglia isolated from E8 to E14 (Fig. 4A), which corresponds to a period when half the neurons are lost due to cell death. PSCA transcripts in ciliary ganglia are barely detectable at E8. However, midway through the cell loss period at E10-12, PSCA levels increase 10-fold. This increase can be attenuated by chronic application of αbtx at a concentration (20µg/day) known to rescue neurons while the α7 nAChR transcript levels double, as might be expected if fewer neurons were dying (Table 1; Bunker and Nishi, 2002). By E14, PSCA expression reaches highest levels, when it is 15 times more abundant than at E8 (Fig. 4A). After E14, which marks the end of cell loss in the ciliary ganglion, there is no further increase in PSCA transcript levels (data not shown). At E14, PSCA transcripts are detected in many neurons, but the relative levels of expression vary from low (Fig. 4B, arrows) to very high (Fig. 4B, arrowheads). Little or no signal is observed with the sense probe (Fig. 4C). Thus, PSCA mRNA correlates with the period of cell loss in the ciliary ganglion and is found in neurons, consistent with a possible role in modulating nicotinic signaling.
To test whether PSCA could modify nicotine-induced responses in ciliary ganglion neurons, we used the retroviral vector RCASBP(A) to express PSCA at E8, an age at which PSCA mRNA is almost undetectable. We removed PSCA-expressing ciliary ganglia and vector-only infected controls at E8, plated the neurons on coverslips, and loaded them with the calcium sensitive dye, Fura-2, in order to quantify intracellular calcium in response to rapidly perfused nicotine. To isolate nicotinic Ca2+ responses from Ca2+ influx due to the opening of voltage-gated ion channels, tetrodotoxin (TTX) and cobalt were added to the perfusion solution (see Materials and Methods).
In controls, nicotine induces a large increase in intracellular calcium, which is partially blocked by the α7 subunit-specific nicotinic cholinergic antagonist, MLA (Fig. 5A). The remaining response is blocked by dihydro β- erythroidin, an antagonist of α3* nAChRs (data not shown). When neurons infected with RCASBP(A)-PSCA are compared to those infected with open-RCASBP(A), there is no difference in the calcium signal induced by depolarization with 25 mM KCl in the absence of TTX and Co2+ (Open = 0.52 ± 0.04 sem; n= 30; PSCA = 0.59 ± 0.06 sem, n = 26), indicating equal loading of Fura-2 and good health of PSCA infected neurons. In neurons overexpressing PSCA, the mean nicotine-induced calcium response is reduced by 43% when compared to open vector-infected cells (Fig. 5B; p<0.001, one-way ANOVA, Bonferroni post test; open: 0.3 ±0.02 sem, n=69; PSCA: 0.17 ±0.01 sem, n=86). This contrasts with the 69% reduction observed when neurons are infected with a tethered αbtx using the same vector and otherwise identical assay conditions (Hruska et al, 2007). The mean response of the PSCA-expressing neurons cannot be significantly lowered by the addition of a solution containing the α7-nAChR specific antagonist, αbtx (Fig 5B). Inclusion of αbtx to open-RCASBP(A) infected neuronal cultures reduces the amplitude of Ca2+ transients to the levels that are comparable to the PSCA infected neurons (Fig. 5B; p<0.001, one-way ANOVA, Bonferroni post test: open: 0.3 ± 0.02, n=69, open+btx: 0.14 ± 0.02, n=31). Thus, retrovirally delivered PSCA expressed at E8 in ciliary ganglion neurons appears to primarily act to suppress activation of α7-nAChRs, but this suppression is less than that observed for tethered αbtx, suggesting that the PSCA affects a subpopulation of neurons or a subpopulation of the nicotinic responses.
Since signaling through α7-nAChRs affects the final number of neurons in the ciliary ganglion, we determined whether the premature expression of PSCA caused a change in neuronal survival. We quantified the total number of neurons (using anti-Islet-1) as well as the number of choroid neurons (using anti-somatostatin) at E14 by using design-based stereology (Lee et al., 2001; Bunker and Nishi, 2002). If infective RCASBP(A)-PSCA particles are injected at St. 8–9 (36 hrs of incubation), virtually all of the ciliary ganglion neurons are infected (Fig. 6A–C); however, if embryos are injected at St 10–13 (48 hrs of development), very few neurons are infected, while many surrounding non-neuronal cells are infected (Fig 6D–F). Little or no infection is detected in the accessory oculomotor nucleus, which innervates the ciliary ganglion (Fig. 6G–I). PSCA expressing ganglia infected at 36 hrs of development contained 35% more choroid neurons than ganglia infected with open-RCASBP(A) (Fig. 7; p<0.0001, one-way ANOVA; open: 4679 ±353.4, n=16; PSCA at 36hrs: 7125 ±355.1, n=13) and this was reflected in a significant change in the total number of neurons in the ganglion (p<0.001 one-way ANOVA; open: 8626 ±418.8, n=16; PSCA at 36hrs: 11173 ±467, n=13). This indicates that the change in the number of neurons expressing somatostatin-like immunoreactivity was not merely due to a shift in neuropeptide expression. Interestingly, the number of ciliary neurons remained unchanged (Fig. 7), suggesting that choroid neurons are more sensitive to the expression of PSCA. When RCASBP(A)-PSCA retrovirus is injected into the neural tube at St .10–13, then PSCA fails to rescue neurons from dying (Fig. 6J; PSCA at 48 hrs: 9352 ±646.1, n=9), despite the fact that neighboring glia are infected and overexpress PSCA.
The principal finding of this study is that PSCA, a molecule originally identified as an antigen upregulated in prostate cancer, is an endogenous prototoxin highly expressed in the nervous system that attenuates signaling through α7 subunit containing nAChRs. PSCA levels correlate with the expression of transcripts encoding α7-nAChRs. In the avian ciliary ganglion, PSCA expression is induced and upregulated between E8 and E14, a time during which half of the neurons are lost by cell death and peripheral synaptogenesis is completed. Furthermore, forcing premature expression of PSCA blocks α7 nAChR activation and prevents choroid neurons from dying. These studies uncover possible modulatory functions for prototoxins during development.
PSCA belongs to the Ly6 superfamily, whose members possess cysteine rich molecules expressed in tissue specific patterns during development and in the adult (Gumley et al., 1995). To date, the function of such molecules has been mysterious because their small size and lack of a transmembrane domain preclude a direct role in mediating cell signaling. Within this family, molecules that fold into a structure homologous to those formed by cobratoxins (Tsetlin, 1999), which bind with high affinity to specific subclasses of nAChRs, have been described as prototoxins and include mouse lynx1 (Miwa et al., 1999), mouse lynx2 (Dessaud et al., 2006), Ly6H (Horie et al., 1998) and SLURP-1 (Chimienti et al., 2003). These endogenous prototoxins are expressed in the CNS and the PNS, and likely act as molecules that interact cell autonomously to modulate nicotinic receptor function in vivo. Indeed, lynx1 and lynx2 coprecipitates with α4/β2 as well as α7 subunit containing nAChRs in mouse CNS and this association alters nAChR kinetics, extent of receptor desensitization and agonist affinity (Miwa et al., 1999; Ibanez-Tallon et al., 2002; Tekinay et al, 2009). Furthermore, neurons from lynx1 null mice exhibit large increases in [Ca2+]i in response to nicotine and, as a result, display age-dependent degeneration that is exacerbated by nicotine and ameliorated by null mutations in nAChRs (Miwa et al., 2006), and lynx2 null mice have altered responses to fear conditioning (Tekinay et al, 2009). Our studies using retroviral mediated expression of avian PSCA in chicken neurons indicate that PSCA exhibits many of the features of a prototoxin: 1) it is cysteine-rich, with a spacing of cysteine residues that is conserved with other members of the family; 2) it is highly expressed in tissues of the nervous system that contain high levels of α7-nAChRs and these data are corroborated by in situ hybridization from mouse brain, where PSCA is detected in Purkinje and granule cell layers of cerebellum, cerebral cortex and hippocampus (http://www.brain-map.org/mouse/brain/Psca.html); 3) it is predicted to form a “three-fingered” tertiary structure similar to that of nicotinic antagonists derived from cobratoxin; 4) it interferes with nicotine-induced increases in [Ca2+]i through α7-nAChRs, but not heteromeric nAChRs. These results suggest that many other members of the Ly6 superfamily may also serve as prototoxins with selectivity for specific classes of nAChRs.
Nicotinic signaling has long been known to play an important role in regulating programmed cell death during development. Blocking neuromuscular transmission with nicotinic antagonists such as d-tubocurarine or α-bungarotoxin is one of the most effective ways to rescue spinal cord motor neurons from dying (Pittman and Oppenheim, 1978; Pittman and Oppenheim, 1979). In the autonomic and some parts of the central nervous system, activation of neuronal nAChRs can directly induce apoptosis. For example, chronic blockade of α7-nAChRs with systemically applied αbtx or MLA prevents cell death of nearly all ciliary ganglion neurons (Meriney et al., 1987; Bunker and Nishi, 2002), although the same is not true for spinal cord motor neurons (Oppenheim et al., 2000). In addition, reduction in Ca2+ influx through α7-nAChRs in a cell-autonomous manner prevents ciliary and choroid neurons from dying, suggesting that large α7-nAChR mediated increases in [Ca2+]i promote cell death of ciliary ganglion neurons during development (Hruska et al, 2007). Immature neurons are especially vulnerable to Ca2+ influx via α7-nAChRs; activation of these channels induces apoptosis of hippocampal progenitor cells but not differentiated hippocampal neurons (Berger et al., 1998). Since the activation of α7-nAChRs can be pro-apoptotic in certain neuronal populations, the signaling through these channels must be precisely regulated.
Our data are consistent with PSCA acting as a modulator of nicotinic signaling in the ciliary ganglion that limits cell death caused by activation of α7-nAChRs in neurons vulnerable to calcium overload. At E8, when PSCA is undetectable, many ciliary ganglion neurons are undergoing apoptotic cell death (Lee, 2001). Subsequent upregulation of PSCA correlates with a significant decrease in cell death, reflected in the stabilization of neuronal cell number. Overexpression of PSCA at E8, when it is not normally expressed, blocks nicotine-induced increases in intracellular calcium through α7-nAChRs and prevents cell death of choroid but not ciliary neurons. Finally, blocking all α7 signaling with αbtx attenuates the developmental increase seen in PSCA, consistent with calcium influx upregulating PSCA as part of a negative feedback loop.
The selective effect of PSCA on the choroid neuron subpopulation contrasts to that of membrane-tethered αbtx, which also blocks nicotine-induced increases in [Ca2+]i via α7- nAChRs (Hruska et al, 2007), but rescues both ciliary and choroid subpopulations. One possible explanation is that tethered αbtx blocks a much higher percentage of the mean nicotine induced response when compared to PSCA (69% versus 43%; Hruska et al, 2007). One likely difference between αbtx and PSCA may be the dissociation constants when bound to α7-nAChRs. The dissociation constant of αbtx is so high that it is virtually irreversible, while this is unlikely to be the case for PSCA. Thus, PSCA may have a differential affinity for choroid neuron nAChRs, or ciliary neuron nAChRs may already be occupied by a prototoxin that does not block calcium influx but cannot be displaced by PSCA. This may be likely because we find two other prototoxin-like molecules, ch3Ly and ch5Ly, in the ciliary ganglion. This differential susceptibility of choroid neurons to PSCA warrants further investigation, but is difficult to pursue because extracellular markers that distinguish ciliary from choroid neurons in acutely dissociated, live preparations have yet to be discovered. Thus, ciliary versus choroid neurons cannot be definitively identified for electrophysiological studies and neither can they be conveniently sorted for biochemical binding or molecular studies. In addition, the measurement of the binding and dissociation constants of membrane-tethered molecules to integral membrane proteins is highly challenging.
The relationship of PSCA expressed in the prostate to its expression in the nervous system is not clear. PSCA was first identified as an antigen enriched in basal cells of the prostate epithelium that is upregulated in high grade, metastatic prostate tumors (Reiter et al., 1998), but undetectable in northern blots of whole human brain. In contrast, when using the more sensitive quantitative PCR on dissected neural tissues, we see significantly higher levels of PSCA than in peripheral organs. Currently, it is unknown whether PSCA modulates α7-nAChR signaling in the prostate or whether it has a completely unrelated function. However, many non-neural cells, such as keratinocytes, lymphocytes, endothelial cells and glia, express α7-nAChRs (reviewed in (Sharma and Vijayaraghavan, 2002). In this context, SLURP1 and 2 of the Ly6, lynx superfamily, regulate apoptosis in keratinocytes (Chimienti et al., 2003; Arredondo et al., 2006; Arredondo et al., 2007). Thus, It is possible that α7-nAChR signaling also regulates cell proliferation in the prostate. Therefore, uncovering the neural and non-neural functions of nAChRs and their accompanying prototoxin modulators will be key for understanding the importance of nicotinic signaling in normal physiology and disease.
We are grateful to Dr Nathanial Heintz at Rockefeller University for providing us with anti-mouse lynx1 antibodies that started this project and for his insights. We would also like to thank Priscilla Kimberly and Sondra Sammut for excellent technical support. The real time PCR and calcium imaging were made possible by the Neuroscience COBRE (5 P20 RR016435). This project was funded by DA17784 to RN.