Spinal neurons require Islet1 for subtype-specific differentiation of electrical excitability

Abstract

Background: In the spinal cord, stereotypic patterns of transcription factor expression uniquely identify neuronal subtypes. These transcription factors function combinatorially to regulate gene expression. Consequently, a single transcription factor may regulate divergent development programs by participation in different combinatorial codes. One such factor, the LIM-homeodomain transcription factor Islet1, is expressed in the vertebrate spinal cord. In mouse, chick and zebrafish, motor and sensory neurons require Islet1 for specification of biochemical and morphological signatures. Little is known, however, about the role that Islet1 might play for development of electrical membrane properties in vertebrates. Here we test for a role of Islet1 in differentiation of excitable membrane properties of zebrafish spinal neurons.

Results: We focus our studies on the role of Islet1 in two populations of early born zebrafish spinal neurons: ventral caudal primary motor neurons (CaPs) and dorsal sensory Rohon-Beard cells (RBs). We take advantage of transgenic lines that express green fluorescent protein (GFP) to identify CaPs, RBs and several classes of interneurons for electrophysiological study. Upon knock-down of Islet1, cells occupying CaP-like and RB-like positions continue to express GFP. With respect to voltage-dependent currents, CaP-like and RB-like neurons have novel repertoires that distinguish them from control CaPs and RBs, and, in some respects, resemble those of neighboring interneurons. The action potentials fired by CaP-like and RB-like neurons also have significantly different properties compared to those elicited from control CaPs and RBs.

Conclusions: Overall, our findings suggest that, for both ventral motor and dorsal sensory neurons, Islet1 directs differentiation programs that ultimately specify electrical membrane as well as morphological properties that act together to sculpt neuron identity.

Figure 1

Caudal primary motor neuron-like neurons have abnormal axonal trajectories that resemble those of interneurons. In all figures, unless indicated otherwise, images present lateral views of embryos (rostral left, dorsal up). (A) In control (Ctl) Tg(mnx1:gfp)ml2 22 to 26 hours post-fertilization (hpf) embryos, primary motor neurons (PMNs) express green fluorescent protein (GFP). The caudal PMN (CaP; large white arrowhead) projects its peripheral axon (small white arrowhead) ventrally. White lines (A, B and C) indicate somite boundaries. (B) In Tg(mnx1:gfp)ml2 E3 morphants, GFP⁺ cells persist in the ventral spinal cord. GFP⁺ CaP-like cells (yellow arrowhead) have somas in CaP positions but project axons centrally rather than peripherally. (C) In another Tg(mnx1:gfp)ml2 E3 morphant, a GFP⁺ CaP-like cell (yellow arrowhead), filled with Alexa 594, has an axon that extends caudally to exit in the neighboring hemisegment (thin yellow arrowhead). (D-L) (D) In uninjected Tg(nrp1a:gfp)js12 embryos, CaP (white arrowhead) has its soma immediately dorsal to the motor axon exit point (white asterisk). (E) E3 morphants have few GFP⁺ ventral neurons. A CaP-like cell (yellow arrowhead) lacks a peripheral axon. (F, F’) GFP⁺ PMNs of control Tg(nrp1a:gfp)js12 embryos (F’, white asterisks) express nrp1a. (G, G’) Following Islet1 knock-down, few nrp1a/GFP⁺ (G’, yellow asterisks) neurons are present. (H-L) In E3 morphants, many GFP⁺ neurons have axons that bypass normal exit points and extend centrally either caudally (H and I), rostrally (J) or in both directions (K). Occasionally, a GFP⁺ CaP-like cell extends a peripheral as well as a central axon (L). (M) The top cartoon depicts control CaP axon morphology, and the six lower cartoons exemplify the range of CaP-like axonal phenotypes revealed by either dye filling or confocal analysis of Tg(nrp1a:gfp)js12 E3 morphants. Scale bars = 50 μm in A (for A to C), D (for D and E), F (for F to G’) and H (for H to L). IC, ipsilateral commissural; KA, Kolmer-Agduhr; VeLD, ventral lateral.

Figure 2

Islet1 knock-down reduces the number of dorsal GFP⁺ neurons in the Tg(-3.4neurog1:gfp)sb4 line. (A-C)Tg(-3.4neurog1:gfp)sb4 24 hours post-fertilization (hpf) Islet1 knock-down reduces the number of dorsal GFP⁺ neurons in the Tg(-3.4neurog1:gfp)sb4 line embryos express green fluorescent protein (GFP) in Rohon-Beard cells (RBs) [38] (white arrow, white outlined cell as example) and dorsal lateral ascending interneurons (DoLAs; asterisk, red circle). (A) In control embryos, central axons (red arrowhead) and peripheral processes (white arrowhead) are GFP⁺. (B) The 5-base mismatched islet1(Sp)E3 MO (CtlMO) has no effect on RB or DoLA morphology. (C) After Islet1 knock-down, central axons (yellow arrowhead) and dorsal cells with RB-like (yellow arrow) or DoLA-like (yellow asterisk) somata remain GFP⁺. However, few peripherally projecting processes are present (white arrowhead in A and B). (D)Tg(-3.4neurog1:gfp)sb4 E3 morphants have 29% less GFP⁺ dorsal neurons versus controls (*P < 0.0001, t-test). GFP⁺ RB neurons were counted in 200 μm spinal cord regions (above yolk sac-yolk sac extension boundary). The number in the bar indicates sample size. (E-G)Tg(mnx1:gfp)ml2 embryos were co-processed for GFP and HNK-1-like immunoreactivities to assess both RB (red) and primary motor neuron (PMN; green) peripheral processes. (E) RB somata (white arrows), central axons and peripheral processes (white arrowhead) are HNK-1 positive. (F) CtlMO has no effect on either motor/interneuron (green) or RB (red) morphology. (G) After Islet1 knock-down, few ventral cells express GFP versus uninjected (Uninj) (E) or control (F) embryos. Few HNK-1⁺ (red) or GFP⁺ (green) projecting peripheral processes are present. (H) The number of DoLA interneurons (twelve hemisegments) were counted in 24 hpf control and morphant Tg(-3.4neurog1:gfp)sb4 embryos. Islet1 knock-down has no effect on the DoLA interneuron number. (I) Commissural primary ascending interneurons (CoPAs) are positive for anti-neurofilament antibody 3A10 staining (asterisk). (J) In 24 hpf embryos, although several spinal neurons are positive for anti-Islet1/2 immunoreactivity (red), CoPAs are not. Scale bars = 50 μm in A (for A to C), E (for E to G) and I (for I and J).

Figure 3

Islet1 knock-down affects sensory neuron differentiation. (A-C) In control embryos (A, B), Rohon-Beard cells (RBs) express runx3. Islet1 knock-down leads to fewer cells with robust expression of runx3 (C). The asterisk denotes a cell expressing the gene. (D-F’)Tg(-3.4neurog1:gfp)sb4 control (D-E’) and E3 morphant (F, F’) 24 hours post-fertilization (hpf) embryos were examined for expression of olig4 (red). (D-E’) In lateral views of the dorsal spinal cord, RNA in situ hybridization reveals expression of the interneuron marker olig4 (red). Green fluorescent protein (GFP)⁺ neurons do not express olig4 and comprise RBs and dorsal lateral ascending interneurons (see Figure  2). (F-F’) Islet1 knock-down leads to an increase in the number of olig4 expressing cells within the dorsal spinal cord. However, similar to RBs of control embryos (D-E’), RB-like neurons do not express detectable levels of olig4(F). (G-I) At 72 hpf, dorsal root ganglia (DRGs) are easily identified as GFP⁺ neurons with large somata located near the spinal cord/notochord border. (G, H) In control embryos, DRG neurons project from their soma bipolar axons that extend dorsally and ventrally (asterisks). (I) Islet1 knock-down reduces the number of GFP⁺ DRGs. Furthermore, for the few GFP⁺ DRGs remaining, their axons show abnormal morphologies. Scale bars = 50 μm in A (for A-C), D (for D-F’) and G (for G-I). CtlMO, 5-base mismatched; islet1(Sp)E3MO; E3MO, E3 morpholino; Uninj, uninjected.

Figure 4

Caudal primary motor neuron-like neurons have voltage-dependent outward current properties that resemble those of ventral lateral descending and Kolmer-Agduhr interneurons rather than caudal primary motor neurons. (A) In uninjected 24 hours post-fertilization (hpf) embryos, caudal primary motor neurons (CaPs), ventral lateral descending neurons (VeLDs), and Kolmer-Agduhr neurons (KA’s, KA”s) have detectable whole-cell inward current (I_Na/Ca). (B) CaPs of control morpholino (CtlMO) injected and uninjected embryos have I_Na/Ca of similar amplitude (A). CaP-like neurons of E3 morphants have I_Na/Ca amplitude that does not differ from control CaPs. (C) I_Na/Ca densities of ventral neurons are not significantly different except for VeLDs versus CaP-like cells (^†P < 0.05, versus VeLD). (D) In control embryos, CaPs have larger amplitude whole-cell outward current (I_Kv/Ca) than do VeLDs, KA’s or KA”s. (E) CaPs of control morphants and uninjected embryos have similar I_Kv/Ca density. In contrast, CaP-like neurons, regardless of the presence or absence of a peripheral axon, have I_Kv/Ca amplitudes that are smaller than those of control CaPs and resemble those of ventral interneurons (D). (F) I_Kv/Ca densities are significantly smaller in VeLDs, KA’s and KA”s compared to control CaPs (*P < 0.001 versus CaP). I_Kv/Ca of CaP-like versus CaPs is significantly smaller (*P < 0.001 versus CaP), but does not differ from VeLDs, KA’s or KA”s. I_Kv/Ca densities do not differ significantly between ventral interneurons. (G) In uninjected embryos, CaPs have larger voltage-dependent potassium current (I_Kv) amplitudes than do ventral interneurons. Outward current properties of KA’ and KA” were not significantly different and are grouped as KA. (H) In control morphants (CaP CtlMO injected), CaP I_Kv resembles that recorded from CaPs in uninjected embryos (G). In E3 morphants, CaP-like I_Kv amplitude is reduced compared to that of control CaPs and more similar to that of ventral interneurons (see G). (I) CaPs have larger I_Kv densities versus interneurons (*P < 0.001 versus CaP). CaP-like neurons have I_Kv densities that are significantly reduced versus CaPs (*P < 0.001 versus CaP) but not interneurons.

Figure 5

In E3 morphants, Rohon-Beard-like cells have electrical properties that resemble those of dorsal lateral ascending interneurons and dorsal commissural interneurons. (A) In uninjected embryos, RB whole-cell inward current (I_Na/Ca) is of larger amplitude than that of dorsal lateral ascending interneurons (DoLA) or dorsal commissural interneurons (Dorsal Comm). (B) In CtlMO injected embryos, RBs have I_Na/Ca amplitude that resembles that of RBs in uninjected embryos (see A). However, I_Na/Ca amplitude of RB-like cells in E3 morphants is smaller than that of control RBs. In contrast, DoLAs in uninjected (A) and E3 morphants have similar inward current amplitudes. (C) RBs of control embryos have significantly larger inward current density than do dorsal interneurons (*P < 0.001 versus RB). The inward current densities of interneurons in control and E3 morphant embryos are not significantly different from each other. However, I_Na/Ca density of RB-like cells is significantly smaller than that of control RBs (*P < 0.001 versus RB) and instead resembles that of DoLAs in control or E3 morphant embryos. In contrast, I_Na/Ca density of RB-like cells is significantly larger than that of Dorsal Comm interneurons (^†P < 0.05 versus Dorsal Comm). (D) In uninjected embryos, whole-cell outward current (I_Kv/Ca) amplitude of RB neurons is larger than that of dorsal interneurons. (E) RB-like neurons have smaller amplitude I_Kv/Ca than do control RBs in uninjected (D) or CtlMO injected embryos (E). In contrast, knock-down of Islet1 has no effect on I_Kv/Ca amplitude recorded from DoLAs (E). (F) Steady-state I_Kv/Ca density of RBs is significantly larger than that of neighboring interneurons (*P < 0.001 versus RB) or of RB-like cells (^†P < 0.05 versus RB) or DoLAs (^#P < 0.01 versus RB) in E3 morphants. Current densities of the DoLA and Dorsal Comm interneurons are not significantly different from each other.

Figure 6

Islet1 knock-down reduces Rohon-Beard cell number and expression of scn8aa in the dorsal spinal cord. (A-C) In 24 hours post-fertilization (hpf) Tg(scn8aa:gfp)ym1 embryos, Rohon-Beard (RB) and RB-like cells express green fluorescent protein (GFP) (dorsal views)[46]. Islet1 knock-down reduces the number of GFP⁺ neurons (C versus A and B). (D) In E3 morphants, the number of GFP⁺ RB-like cells is 44% that of GFP⁺ RB cells in controls (*P < 0.001) (see Methods). (E-G) In uninjected (Uninj) and control 24 hpf double Tg(scn8aa:gfp,-3.4neurog1:dsRed) embryos, most RB neurons express both reporters (E, F) and few express only dsRed (arrowheads) (lateral views). In E3 morphants, many RB-like cells express only dsRed (G, arrowheads). (H) Compared to RBs, more RB-like cells express only dsRed (^#P < 0.01). (I-K) Following Islet1 knock-down (K), less scn8aa mRNA (asterisks) is detected compared to controls (I) (lateral views). (J)The sense control scn8aa probe reveals little signal, indicating specificity of the antisense probe (I and K). (L-N) In 24 hpf double transgenic Tg(ssx-mini-ICP:egfp,-3.4neurog1:dsRed) embryos, most RB cells (asterisks) express both reporter proteins (N). Fewer ventral interneurons express GFP (L) versus dsRed (M) (arrowheads). (O) Quantitative reverse transcription PCR analysis (see Methods) shows a five-fold reduction in scn8aa expression in RB-like cells of E3 morphants (*P < 0.04). (P, Q) RBs express vglut2.1/2.2 mRNA (red) (lateral views). In uninjected and CtlMO injected embryos, vglut2.1/2.2 mRNA (red) colocalizes (white arrows) with GFP⁺ RBs (green). (R) In E3 morphants, vglut2.1/2.2 mRNA (red) co-localizes with GFP⁺ RB-like neurons (green; yellow arrow), indicating no obvious effect of Islet1 knock-down. (S, T) GABAergic neurons express gad65/67 mRNA (red). (U) Islet1 knock-down has no obvious effect on gad65/67 in the dorsal domain. Scale bars = 50 μm in A (for A to C), E (for E to G), I (for I to K), L (for L to N), P (for P to R) and S for (S to U).

Figure 7

Caudal primary motor neuron and caudal primary motor neuron-like neurons fire action potentials with different properties. (A, B) Action potentials were evoked from caudal primary motor (CaPs), ventral lateral descending (VeLDs) and CaP-like neurons (see Methods). The insets align action potentials at their peaks to highlight kinetic differences. (A) During an action potential, the membrane potential repolarizes faster in CaPs than it does in VeLDs. (B) For CaP action potentials, the 5-base mismatched islet1(Sp) E3 morpholino (CtlMO) has no obvious effect on repolarization. In comparison, CaP-like action potentials of E3 morphants repolarize slowly. (C) Kolmer-Agduhr neurons (KAs) do not fire action potentials in response to brief depolarizing current injections. (D) After prolonged injection of hyperpolarizing current, KA”s fire regenerative responses. A single regenerative response recorded from another KA” is enlarged (right). (E) KA” action potentials have slower rates of rise than those elicited from CaPs (*P < 0.001 versus CaP) or VeLDs (^#P < 0.01 versus VeLD). However, the rate of rise of CaP-like action potentials is not significantly different from that of KA”s, VeLDs or CaPs. In contrast, the rate of decay of CaP action potentials is significantly faster than that of VeLD (^†P < 0.01 versus CaP), KA”s (*P < 0.001 versus CaP) or CaP-like action potentials (^P < 0.01 versus CaP). (F) Regenerative responses fired by KA”s have prolonged rise (^#P < 0.01 versus CaP or CaP E3MO) and decay (*P < 0.001 versus CaP) times. VeLDs fire action potentials of significantly longer duration than do CaPs (^†P < 0.05 versus CaP). Similarly, KA” regenerative responses have significantly longer durations than do those elicited from CaPs (*P < 0.001 versus CaP) or CaP-like cells (^†P < 0.05 versus CaP E3MO). (G) CaPs, VeLDs and CaP-like neurons have similar input resistance. Compared to CaP, KA”s have significantly greater input resistance (*P < 0.001 versus CaP) or CaP-like cells (^#P < 0.01 versus CaP E3MO).

Figure 8

Rohon-Beard and Rohon-Beard-like cells fire action potentials with different properties. (A, B) Action potentials were evoked from dorsal spinal neurons in 24 hours post-fertilization (hpf) embryos (see Methods). (A) Rohon-Beard (RB) action potentials waveform have a distinctive overshoot and afterhyperpolarization (AHP). (B) RB-like cells of E3 morphants fire action potentials without a prominent AHP. (C) Aligning the peaks of RB and RB-like action potentials highlights kinetic differences. (D, E) Several properties of action potentials (rate of rise, rise time, rate of decay, decay time, duration) were evaluated at rheobase. (D) RBs fire action potentials with faster rates of rise and decay than those elicited from dorsal lateral ascending interneurons (DoLAs) and dorsal commissural interneurons (Dorsal Comms) (*P < 0.001 versus RB). Compared to RBs, RB-like cells fire action potentials with slower rates of rise (^#P < 0.01 versus RB) and decay (^†P < 0.05). Compared to DoLAs and Dorsal Comms, RB-like cells fire action potentials with faster rates of rise (^P < 0.01 versus DoLAs and Dorsal Comms) and decay (^•P < 0.01 versus DoLA; °P < 0.05 versus Dorsal Comm). (E) Compared to RBs, DoLAs fire action potentials with increased rise time, prolonged decay time and longer duration (*P < 0.001 versus RBs). Dorsal Comms fire action potentials with longer decay time (^#P < 0.01) and duration (^†P < 0.05) than do RBs. DoLAs and Dorsal Comms fire impulses that have significantly different durations (^#P < 0.01 versus DoLA). RB-like cells fire impulses with decreased rise times and briefer decay time than do DoLAs (^P < 0.001 versus DoLAs). (F) RB-like cells fire impulses with small AHP amplitudes in contrast to those of RBs (*P < 0.001) or DoLAs (^†P < 0.05). (G) The input resistances of RB-like and RB cells are not significantly different but significantly lower than those of DoLAs and Dorsal Comms (*P < 0.001).