Zebrafish (Danio rerio) adults were bred according to guidelines outlined in The Zebrafish Book 20. Embryos were incubated at 28.5°C in embryo medium (130 mM NaCl, 0.5 mM KCl, 0.02 mM Na₂HPO₄, 0.04 mM KH₂PO₄, 1.3 mM CaCl₂, 1.0 mM MgSO₄, 0.4 mM NaH₂CO₃) and staged according to external morphology 21.

Whole cell voltage clamp recordings were obtained from zebrafish spinal cord RBs as previously described 111822. Voltage clamp recordings from caudal primary motoneurons (CaPs) were obtained from the Tg(hb9:GFP) line (gift of Drs Michael Fox and Joshua Sanes, Harvard University, Cambridge, MA, USA) that express green fluorescent protein (GFP) in motoneurons 23. Tg(hb9:GFP) zebrafish were immobilized in Ringer solution (145 mM NaCl, 3 mM KCl, 1.8 mM CaCl₂, and 10 mM HEPES, pH 7.2) containing 0.02% tricaine (Sigma St Louis, Missouri, USA) and glued laterally to glass coverslips. Glass dissecting needles sufficed for removal of skin and detachment of overlying muscle fibers. Muscle fibers and secondary motoneurons were removed by a suction pipette to expose primary motoneurons. Three properties identified CaPs: GFP expression, cell body size (approximately 10 μM diameter), and ventrally projecting axons 24. For initial experiments, we used a reduced extracellular sodium bath solution (30 mM NaCl, 97 mM N-methyl glucamine, 20 mM tetraethylammonium (TEA), 3 mM KCl, 2 mM CoCl₂, and 10 mM HEPES) to reduce potential series resistance voltage errors arising from large I_Na amplitudes. However, some experimental manipulations (for example, knockdown of sodium channel α-subunits or phosphatase blockade) reduced I_Na amplitudes; in these cases, we used a normal 125 mM extracellular sodium concentration to increase I_Na amplitudes and the sensitivity of our measurements. Glass electrodes (2.0 to 3.5 MÙ) were filled with solution containing 10 mM NaCl, 135 mM CsCl, 10 mM EGTA, and 10 mM HEPES. We subtracted passive leak currents and capacitive transients from recordings of voltage-gated sodium using a P/8 protocol. Data were acquired using an Axopatch 200B amplifier (Axon Instruments, Foster City, California, USA) and analyzed with Clampfit8 (Axon Instruments) and Origin software (OriginLab, Northampton, Massachusetts, USA).

Results are presented as means ± standard errors. Statistical analysis was performed with Origin v7.0 software (OriginLab). Statistical comparisons of means were performed by one-way ANOVAs with Bonferroni corrections for multiple comparisons.

T4 (3,3',5,5'-tetraiodo-L-thyronine (thyroxine); Sigma) was prepared as a 30 mM stock solution in dimethyl sulfoxide (DMSO) that was diluted to final concentrations in extracellular recording solution immediately before use. Vehicle (DMSO) control experiments indicated that the final concentration of DMSO (0.001%) had no effect on I_Na amplitudes; therefore, control and vehicle control data were pooled (Control/DMSO). Kinase and phosphatase inhibitors were applied to neurons in semi-intact preparations of the zebrafish embryo after obtaining control recordings prior to treatment. PD98059 (50 μM; Sigma), 1 μM SB203580 (Sigma), or okadaic acid (OA; 1 nM to 1 μM; Sigma) was applied for 1 hour at room temperature before obtaining post-treatment recordings. The drugs remained in the bath during post-treatment recording.

Whole mount embryos (24 to 48 hpf) were processed for immunocytochemistry as previously described 25. The primary antibody, mouse anti-human monoclonal LM609 (Millipore, Billerica, MA, USA), was diluted 1:100. Secondary antibody was applied overnight at 4°C (1:500; goat anti-mouse conjugated to Alexa 568; Invitrogen-Molecular Probes, Carlsbad, California, USA). Controls consisted of experiments done with LM609 that had been previously incubated with 50 μg/ml human αVβ3 (Millipore). In some experiments, the Tg(isl3:GFP) line (gift of Drs Andrew Pittman and Chi-Bin Chien, University of Utah) expressing GFP in RBs or the Tg(hb9:GFP) line expressing GFP in motoneurons were used. GFP expression was revealed using a rabbit anti-GFP antibody conjugated to Alexa 488 (1:400; Invitrogen). For analysis, embryos were mounted in a 1% low melting point agarose solution and imaged using a Zeiss Pascal Confocal Microscope using 10× or 40× objectives and separate 488 and 568 laser lines. Fluorescent images were collected digitally as z-stacks of 2 μm slices. Data are presented as projections of 20 to 25 slices.

Antisense oligonucleotide morpholinos (MOs) targeting sodium channels Na_v1.6a (1.6 MO) and Na_v1.1l (1.1 MO) were synthesized and prepared as previously described 18. Injection solutions contained the dye Fast Green (1%) to report efficient delivery of the MO to animal cells. For each Na_v1 MO, control MOs were synthesized by introducing mismatches at five positions. Embryos that had either Na_v1.6a or Na_v1.1l sodium channel subunit knock-down were created by injection of 2 to 3 hpf wild-type embryos with solution containing 0.3 mM MO antisense oligonucleotide 18. All MOs have been used previously and tested by standard control experiments 181926.

Embryos that had Fast Green within the animal cell 15 minutes post-injection were transferred to a petri dish containing embryo medium (130 mM NaCl, 0.5 mM KCl, 0.02 mM Na₂HPO₄, 0.04 mM KH₂PO₄, 1.3 mM CaCl₂, 1.0 mM MgSO₄, 0.4 mM NaH₂CO₃) and then raised at 28°C until 48 hpf. Embryos that were injected with the 1.6 MO or 1.1 MO and selected for recording are referred to as Na_v1.6a or Na_v1.1l morphants, respectively.

In other systems, rapid thyroid hormone signaling involves intracellular signaling kinase pathways such as MAPK (ERK1/2) and MAPK (p38) 71213272829. To identify intracellular mediators of rapid T4 signaling in RBs, we used a pharmacological approach. We inhibited MAPK (ERK1/2) or MAPK (p38) signaling by using the blockers PD98059 or SB203580, respectively. In addition, because phosphatase effects oppose kinase action we used OA to block serine/threonine phosphatases. At low OA concentrations we blocked PP2A (1 to 20 nM OA) and at higher concentrations we blocked both PP2A and PP1 (1 μM OA). After incubation of spinal cord preparations in conditions of kinase or phosphatase blockade, we tested for effects of kinase and phosphatase inhibitors on I_Na amplitude, in the absence and presence of T4.

Control/DMSO cells displayed a I_Na peak amplitude of 1,665 ± 116 pA (n = 23; Figure 1D) and responded to acute T4 application with a 39 ± 5% increase in amplitude (n = 4; P < 0.05). To test for involvement of ERK1/2, we used PD98059 (50 μM), a specific inhibitor of the ERK 1/2 pathway component MEK1 3031. By itself, PD98059 did not significantly alter I_Na peak density (1,756 ± 217 pA; n = 16; P = 0.69) compared to control/DMSO (Figure 1D). Further, RBs exposed to PD98059 could still respond to T4 by rapidly increasing I_Na amplitude (Figure 2A, C). Compared to control/DMSO cells, however, PD98059-treated cells showed a blunted response to T4 (Figure 2C).

Whereas the PD98059 results suggest that the ERK1/2 pathway may partially mediate the rapid effects of T4, the data do not implicate ERK1/2 signaling in tonic regulation of the number of available sodium channels. In contrast, nongenomic T4 signaling regulates both the rapid response to T4 as well as the tonic levels of available sodium channels in RBs 11. Overall, the PD98059 results suggest minimal involvement of the ERK1/2 pathway in T4 regulation of RB I_Na.

We next tested the contribution of the MAPK (p38) pathway that mediates rapid thyroid hormone signaling triggered by T3 13 and regulates I_Na density 16. In contrast to blockade of the ERK1/2 pathway, pharmacological inhibition of the p38 pathway increased tonic I_Na amplitudes (in the absence of exogenous T4) 1.48-fold (2458 ± 134 pA; n = 14; P < 0.05, ANOVA) (Figure 1D). Moreover, following SB203580 treatment, T4 no longer produced a rapid increase in I_Na amplitude. In fact, after SB203580 treatment, T4 application led to a 22 ± 4% (n = 5) decrease in RB I_Na amplitude (Figure 2A, C). Overall, the effects of SB203580 support involvement of the p38 pathway in regulation of both the resting levels of available RB sodium channels as well as the rapid response of RB I_Na to T4.

The SB203580 results suggest that T4 acts rapidly on RBs by opposing ongoing MAPK p38 signaling either by inhibiting the kinase or by activating relevant serine/threonine phosphatases. Two serine/threonine phosphatases, PP2A and PP1, are both ubiquitously expressed in the zebrafish spinal cord at 48 hpf 17, a time when T4 rapidly regulates RB sodium current. To test for involvement of PP2A or PP1, we incubated zebrafish spinal cords with the serine/threonine phosphatase inhibitor OA prior to recording I_Na from RBs. At concentrations of 1 to 20 nM, OA specifically inhibits PP2A. However, 1 and 20 nM OA did not significantly alter I_Na peak amplitudes in the absence of T4 (1,761 ± 179 pA (n = 6) and 1,798 ± 283 pA (n = 11), respectively). In contrast, the IC₅₀ for PP1 inhibition by OA is much higher (approximately 0.5 μM) 32. OA at 1 μM produced a drastic 90% reduction in resting RB I_Na peak amplitudes (169 ± 33 pA; n = 7; P < 0.00001, ANOVA; Figure 1D). Further, following 1 μM OA treatment, T4 no longer increased RB I_Na amplitude (-6 ± 4% change; n = 7; P = 0.98 versus no T4 added; Figure 2A, C). Both the tonic reduction in I_Na peak amplitude and the occlusion of rapid T4 effects under conditions of PP1 blockade support the view that PP1 activity modulates RB I_Na amplitudes.

We tested whether T4's rapid action on RB I_Na amplitude targeted a specific voltage-gated sodium channel isotype. RBs express two different sodium channel α-subunit genes, scn1.1l and scn8aa, which code for the voltage-gated sodium channel proteins Na_v1.1l and Na_v1.6a, respectively 33. Interestingly, the conductance carried by the mammalian homologue of Na_v1.6a is significantly reduced by activation of MAPK p38 16. We knocked down either Na_v1.1l or Na_v1.6a using MOs, as done previously for effective and selective elimination of specific Na_v1 proteins in the zebrafish embryo 1819.

As found previously, knockdown of either Na_v1.6a or Na_v1.1l α-subunits led to decreased RB I_Na peak amplitudes (Figure 3D) 18. Further, injection of the control Na_v1.6a 5-missense MO did not significantly alter RB I_Na peak density. We next tested whether knock-down of either sodium channel subunit prevented T4's rapid modulation of RB I_Na by acute application of T4 to morphant embryos. We found that T4 increased RB I_Na amplitude in control morphants (45 ± 8%; P < 0.05 versus no hormone) to a similar extent as in wild-type embryos (39 ± 5%) 11 (Figure 3A, E). However, in Na_v1.6a morphants, T4 application did not increase I_Na amplitudes (-16 ± 5%; P < 0.05 versus control morphants), indicating that rapid T4 action targets Na_v1.6a channels.

In contrast, in Na_v1.1l morphant embryos, T4 application increased RB I_Na (63 ± 10%; P < 0.05 versus no hormone), suggesting that T4 does not require Na_v1.1l channels to increase RB I_Na density. In Na_v1.1l morphants, the increase in I_Na amplitude induced by T4 actually exceeded that produced in controls (63 ± 10% versus 45 ± 8%; P < 0.05), presumably because Na_v1.6a channels carry the majority if not all of the remaining current 18. The effects of T4 on Na_v1.6a and Na_v1.1la morphants indicate that rapid T4 signaling targets Na_v1.6a sodium channels.

Antagonists of thyroid hormone (3,3',5,5'-tetraiodothyroacetic acid (tetrac)) or of the integrin αVβ3 block rapid T4 signaling in RBs 11. We next tested whether the effects of thyroid hormone antagonism or αVβ3 blockade targeted Na_v1.6a-mediated current. We exposed Na_v1.6a morphants to the thyroid hormone analog tetrac, or the αVβ3 function blocking antibody LM609. To more readily detect changes in RB I_Na amplitude in Na_v1.6a morphants, we raised the extracellular sodium concentration to 125 mM to increase I_Na amplitudes. We previously showed that tetrac reduces RB I_Na amplitudes in wild-type embryos 11. However, in Na_v1.6a morphants, tetrac did not significantly change RB I_Na amplitudes compared to Na_v1.6a morphants unexposed to tetrac (Figure 3F). We had also previously demonstrated that LM609 reduced RB I_Na amplitudes by 46% in wild-type embryos 11. In contrast, in Na_v1.6a morphants, LM609 injection did not significantly affect I_Na peak amplitudes compared to uninjected Na_v1.6a morphants (Figure 3F). The lack of effect of either tetrac or LM609 in Na_v1.6a morphants further supports that the rapid T4-integrin signaling pathway specifically targets Na_v1.6a channels.

Whether T4 induced increases in sodium current occur throughout development or if T4 signaling begins at a defined developmental stage is unknown. The above results combined with our previous study 11 indicate that in order to respond rapidly to T4 at 48 hpf, RBs require integrin αVβ3 and the Na_v1.6a sodium channel α-subunit. We next determined whether αVβ3 is present and if RBs respond rapidly to T4 at earlier stages. We reasoned that absence of integrin αVβ3 would prevent T4 from rapidly modulating RB I_Na. Accordingly, we tested our prediction by determining the spatial and temporal αVβ3 expression pattern.

To determine when RBs express the plasma membrane T4 receptor, we performed immunocytochemistry using the LM609 antibody that specifically detects the αVβ3 dimer 34. As expected, at 48 hpf the LM609 antibody revealed immunoreactivity in dorsal spinal cord cells (Figure 4A) and cells in the ventral spinal cord. However, at 24 hpf, no immunoreactivity was detected (Figure 4B). To test whether immunoreactivity was specific for αVβ3, we pre-incubated LM609 with integrin αVβ3 protein prior to zebrafish application. We found pre-incubation of LM609 with αVβ3 protein prevented detection of immunoreactivity (Figure 4C), indicating specificity of immunostains for αVβ3.

LM609 immunoreactive cells localized to the dorsal spinal cord where RBs reside. To identify LM609 immunoreactive cells as RBs, we used transgenic Tg(isl3:GFP) embryos. In this line, the isl3 promoter drives GFP expression in RBs (A Pittmann and Chi-Bin Chien, personal communication). In 48 hpf Tg(isl3:GFP) embryos (Figure 4D–F), LM609 immunoreactivity colocalized with GFP, revealing αVβ3 expression on RB bodies. This result is consistent with rapid αVβ3-dependent T4 signaling in 48 hpf RBs 11. In contrast, at 24 hpf, zebrafish embryos did not show specific LM609 immunolabeling in either the dorsal or ventral spinal cord (Figure 4B). These data indicate that αVβ3 dimers appear on RBs after 24 hpf.

At 24 hpf, Na_v1.6a underlies a portion of RB I_Na 18. Nonetheless, according to our model, the lack of αVβ3 expression on RBs at 24 hpf would prevent T4 from rapidly modulating RB I_Na To test this prediction, we acutely applied 30 nM T4 to the spinal cords of 24 hpf embryos and recorded RB I_Na. In contrast to results obtained from 48 hpf embryos 11, we found that T4 had no significant effect on RB I_Na amplitude at 24 hpf (Figure 5A, B). These results indicate that acute modulation of RB I_Na requires αVβ3.

As another test of the requirement for integrin αVβ3, MAPK p38/PP1, and the Na_v1.6a sodium channel α-subunit for rapid T4 signaling, we sought to identify another neuronal cell type that expressed these critical components and test whether T4 could also rapidly modulate I_Na amplitude. The intracellular messengers are ubiquitously expressed in the zebrafish spinal cord 17, and several ventral spinal cord neurons, including interneurons and motoneurons, express Na_v1.6a 1933. We detected LM609 immunoreactivity in the ventral spinal cord (Figure 4A) and now determined the identity of these neurons by using the Tg(hb9:GFP) line. In Tg(hb9:GFP) embryos, motoneurons express GFP, allowing morphological assessment of cell bodies and axonal projections 23. At 48 hpf, a subset of GFP expressing cells in Tg(hb9:GFP) embryos were also immunoreactive for LM609 (Figure 6A–C). In particular, ventral neurons with large diameter cell bodies, a hallmark of primary motoneurons, were co-positive for GFP and LM609 (Figure 6B, C).

At 48 hpf, the spinal cord contains three different types of primary motoneurons 24. One primary motoneuron, CaP, expresses Na_v1.6a at 48 hpf 1933. Because rapid T4 modulation of I_Na targets Na_v1.6a, co-expression of Na_v1.6a and integrin αVβ3 raised the possibility that CaPs might respond to T4 with a rapid increase in I_Na amplitude. To test this possibility, we applied 30 nM T4 to CaP motoneurons, identified in Tg(hb9:GFP) 48-hpf embryos by cell body size and ventrally projecting axons, while recording I_Na. In control CaP recordings, I_Na peak density decreased over 5 minutes by 7 ± 2%. However, CaP I_Na density significantly increased by 28 ± 8% (P < 0.005) after acute T4 application (Figure 6D–F). These results support our model that rapid regulation of I_Na density by T4 requires αVβ3 and targets Na_v1.6a channels.

Our results identify messengers and targets of a rapid thyroid hormone signaling pathway that functions in the zebrafish embryonic nervous system. The T4 pathway rapidly induces increased sodium current amplitudes and requires the sodium channel isotype Na_v1.6a even though neurons express several different sodium channel isotypes. Moreover, the results suggest that the phosphorylation state of an involved protein, perhaps even the targeted sodium channel isotype, determines sodium channel activity.

We propose a model in which the phosphorylation status of a particular protein or set of proteins regulates RB I_Na amplitude, and T4 rapidly alters the phosphorylation status of relevant protein(s). Further, the data suggest that serine/threonine phosphorylation and dephosphorylation reduce and increase, respectively, RB I_Na amplitude. To increase RB I_Na amplitude, T4 may rapidly activate PP1 and/or inhibit p38. Consistent with our findings, PP1 modulates I_Na amplitudes in rat striatal neurons 35, which express Na_v1.6 36. Schiffmann et al. 35 found that PP1 blockade reduced rat striatal I_Na amplitudes, similar to our results in zebrafish RBs.

Our data do not provide information about the identity of the phosphorylated protein(s). One possibility is that T4 binding to αVβ3 activates intracellular pathways that directly phosphorylate sodium channel α-subunits. In ND7/23 cells transfected with mammalian Na_v1.6 sodium channels, activation of p38 produces a decrease in I_Na 16. Of particular relevance, biochemical analysis demonstrated that p38 activity regulated phosphorylation of a specific Na_v1.6 serine, S553, revealing the Na_v1.6 sodium channel as a direct p38 phosphorylation target 16. On this basis, the RB T4-αVβ3 pathway may modulate I_Na by regulating the phosphorylation state of the conserved serine residue in zebrafish Na_v1.6a.

One result that the model does not fully account for, however, is the large reduction in RB I_Na amplitude produced by PP1 inhibition. We previously reported that either T4 or αVβ3 blockade reduced I_Na by only 50%, yet 1 μM OA reduced RB I_Na by nearly 90%. This discrepancy could be attributed to different degrees of PP1 inhibition by 1 μM OA versus T4/αVβ3 blockade. The large decrease in I_Na amplitude produced by 1 μM OA could also reflect phosphorylation effects triggered by non-T4-dependent mechanisms. For example, protein kinases C and A also have effects on Na_v1.6 amplitude 37.

The important developmental roles of integrins as cell surface adhesion proteins have been well studied 8. Less-well studied, however, is the potential role of integrins as receptors for hormones. Because neurons and glia 38 express αVβ3, nongenomic T4 signaling via αVβ3 may play an important role in nervous system development.

Here, we focused on integrin's role as a plasma membrane receptor for thyroid hormones that traditionally signal through nuclear receptors. We focused on the integrin dimer, αVβ3, a protein that is important for neuronal migration and axon extension 394041, and is expressed on dorsal root ganglia 4243. The fact that RGD (Asp-Gly-Arg) proteins block rapid T4 signaling mediated by αVβ3 7 suggests that the hormone interacts with the RGD recognition site 44. Davis et al. suggested that, in addition to αVβ3, seven other RGD integrin dimers may function as thyroid hormone receptors 28.

In addition to T4, the iodothyronine T3 binds to integrin αVβ3 and activates both ERK1/2 and phosphatidyl inositol 3-kinase 45. If T3 interacts with RB αVβ3 to activate ERK1/2 and phosphatidyl inositol 3-kinase, our data indicate that activation of these signaling pathways has no effect on RB I_Na amplitude. Although we found that T3 does not affect RB I_Na, previous studies show T3 can increase I_Na depending on cell type. For example, chronic T3 application increases I_Na in cultured rat hippocampal neurons 46, but not rat cortex in vitro. This result was attributed to increased nuclear thyroid hormone receptor expression in hippocampus versus cortex leading to differential genomic regulation of sodium channel expression. Additionally, T3 rapidly increases I_Na in cultured myocytes 4748 through mechanisms that involve rises in intracellular calcium 48 and protein kinase C 47. Altogether, both T4 and T3 nongenomic signaling result in a variety of downstream consequences due to the diversity of signaling mechanisms activated by plasma membrane thyroid hormone receptors.

During intrauterine stages, the human embryo requires maternally provided thyroid hormone for normal development 495051. However, the specific roles and underlying mechanisms of thyroid hormone action during embryogenesis are poorly understood. Thyroid hormone signals nongenomically to regulate migration of neural cells in the embryonic nervous system 52. Our data indicate that thyroid hormone, acting rapidly via a plasma membrane receptor, shapes emerging properties of neuronal excitability. Specifically, thyroid hormone rapidly modulates neuronal sodium current by targeting the Na_v1.6a subunit. The implications for mammals are substantial because the mammalian homologue, Na_v1.6, shows widespread expression in both the central and peripheral nervous systems 5354 and is highly expressed during embryonic stages 55. Moreover, Na_v1.6a plays important developmental roles for sensory neuron survival and motoneuron axon growth in zebrafish 1925. Taken together, these findings indicate that modulation of Na_v1.6 current during embryonic stages serves as a strategic way to regulate both structural and functional development of the nervous system.

In the context of our results, periods of thyroid hormone deprivation during development would decrease sodium current in neurons expressing both αVβ3 and Na_v1.6, leading to reduced excitability. Conversely, an excess of thyroid hormone would increase sodium current and potentially induce pathological hyperexcitability, associated with seizures and developmental abnormalities. Interestingly, increased expression of Na_v1.6 channels are associated with epileptogenesis in mouse hippocampal neurons through mechanisms of enhanced excitability 56, and acute increases in T4 have been reported to cause seizures in humans 57. Also of note, mice without the functional sodium channel gene SCN8A are somewhat resistant to seizures 58 and children with congenital hypothyroidism have a significantly reduced incidence of febrile convulsions 59. Whether thyroid hormone can acutely influence seizure activity through αVβ3-dependent regulation of Na_v1.6 in mammals warrants further study. Altogether, T4 regulation of Na_v1.6a current provides an important mechanism to influence neuronal activity and development.

We focused on rapid T4 signaling in the embryonic nervous system. However, αVβ3 and Na_v1.6 are also present in the adult nervous system, raising the possibility that T4 acutely regulates sodium current in adults. During adult stages, Na_v1.6 is the primary sodium channel isoform expressed at nodes of Ranvier 53. T4 induced modulation of Na_v1.6 mediated current would alter I_Na at nodes of Ranvier and, therefore, regulate axonal conductance. The mechanism of thyroid hormone action on adult neurons is unclear, yet alterations in Na_v1.6 current could result in deficits in sensory neuron axonal conductance and could account for states of hyper- or hypo-reflexia observed in hyper- or hypothyroid patients, respectively 6061. Studies on T4-αVβ3's activation of angiogenesis and tumor cell proliferation also have clinical corollaries in adults as hypothyroid states reduce tumor cell proliferation in gliomas 62.